Periodic Reporting for period 2 - InDEStruct (Integrated Design of Engineering Structures)
Periodo di rendicontazione: 2020-04-01 al 2022-09-30
Doctoral training, at the core of MSCA-ITN programmes, aims to train engineers with high quality technical and transferable skills. The main driver for the InDEStruct project was the distinct lack of well-trained scientists in the area of multi-disciplinary optimisation of engineering systems that involve thermal and mechanical loading caused by vibration, aerothermal flow and their combined effect.
Industry 4.0 is about digital transformation of companies. One concept exploiting this for mechanical design and manufacture is the idea of a digital twin. It is a virtual representation of a physical system serving as a real-time digital surrogate, providing an indication of the system’s performance. Data is sent to the digital model from the physical product with the former then able to provide performance measures in order to optimize the product. This increases company productivity, as it can reduce or replace expensive and time-consuming physical testing of many candidate designs.
A digital twin is necessarily a multi-disciplinary concept, where well-trained scientists integrate technology that may arise from technical fields that could pose competing design constraints to provide indicators of engineering performance.
Vestas aircoil A/S, part of the InDEStruct consortium, designs and manufactures heat exchangers for marine and locomotive engines. A digital twin can provide warning concerning the health of the cooler, such that the risk of a train breaking down inside a tunnel can be minimized, thereby also contributing to human well-being. Such a digital twin required fundamental multi-disciplinary understanding of systems produced by this project. One study focussed on how vibrations and stress response will influence the life of the materials used for coolers. In this case one of the production methods of tomorrow: added metal manufacturing. To predict the vibration characteristics an efficient and accurate design tool was developed in parallel. Both studies were then complemented by an experimental modal analysis study of tube and fin structures, used in heat exchangers, and incorporating the cooler response to vibration and identify response changes, which can be an indicator of failure. Vibrations alongside corrosion being primary reason for charge air cooler failures. The final study focussed on heat exchanger optimisation, combining all material and dynamic aspects which included a vibration model/constraint with aerothermal modelling.
Industry 4.0 is about digital transformation of companies. One concept exploiting this for mechanical design and manufacture is the idea of a digital twin. It is a virtual representation of a physical system serving as a real-time digital surrogate, providing an indication of the system’s performance. Data is sent to the digital model from the physical product with the former then able to provide performance measures in order to optimize the product. This increases company productivity, as it can reduce or replace expensive and time-consuming physical testing of many candidate designs.
A digital twin is necessarily a multi-disciplinary concept, where well-trained scientists integrate technology that may arise from technical fields that could pose competing design constraints to provide indicators of engineering performance.
Vestas aircoil A/S, part of the InDEStruct consortium, designs and manufactures heat exchangers for marine and locomotive engines. A digital twin can provide warning concerning the health of the cooler, such that the risk of a train breaking down inside a tunnel can be minimized, thereby also contributing to human well-being. Such a digital twin required fundamental multi-disciplinary understanding of systems produced by this project. One study focussed on how vibrations and stress response will influence the life of the materials used for coolers. In this case one of the production methods of tomorrow: added metal manufacturing. To predict the vibration characteristics an efficient and accurate design tool was developed in parallel. Both studies were then complemented by an experimental modal analysis study of tube and fin structures, used in heat exchangers, and incorporating the cooler response to vibration and identify response changes, which can be an indicator of failure. Vibrations alongside corrosion being primary reason for charge air cooler failures. The final study focussed on heat exchanger optimisation, combining all material and dynamic aspects which included a vibration model/constraint with aerothermal modelling.
For multi-disciplinary systems optimisation, the four Early Stage Researchers (ESRs) developed tools and methods to improve the design of a heat exchanger. The four fellows performed analyses and laboratory experiments from the perspective of their individual project aims, focussing on one specific technical aspect. In addition, ESR1 drew together these pieces of technical information within an integrated design framework. This serves as a generic design process highlighting the challenges and benefit of a multi-disciplinary approach. ESR1 completed an automated parametric design of a helical tube and its fluid flow analysis (CFD). This considered a charge air cooler optimization problem, which included constraints on pressure, size, weight, heat dissipation and the newly developed dynamic model from ESR2. ESR2 developed computationally efficient models of modal vibration of heat exchanger components; directly related to in-situ experimental modal and wave analysis of heat exchanger components and their assembly conducted by ESR3. ESR4 performed a fatigue evaluation of additively manufactured 316L stainless steel for said heat exchanger, with attention to crack propagation and build direction in a 3D printer. Together ESR4 and ESR2 evaluated the fatigue life of an exemplar structure.
These results have shown the benefit of a multi-disciplinary approach to optimization and design development. All these tools and outcomes have been documented and transferred into company practice within the design and validation teams. It has provided valuable input to the future project directions that the company will take and the challenges are on the agenda for the R&D department in the future. In addition, the original research and outcomes arising from this study have commenced generating peer reviewed research publications.
Summarising the main results, the design process is driven by geometry generation, calculation of heat transfer & flow, structural analysis, vibration and fatigue testing; the ESR team worked towards an integration of these for the benefit of the lifetime of heat exchangers. The practical lesson learnt is that there is a tendency for technical specialists to “work in isolation” until brought together by the design constraints. This highlights the industrial need to have specialist technology roles integrated within design teams possessing good communication.
These results have shown the benefit of a multi-disciplinary approach to optimization and design development. All these tools and outcomes have been documented and transferred into company practice within the design and validation teams. It has provided valuable input to the future project directions that the company will take and the challenges are on the agenda for the R&D department in the future. In addition, the original research and outcomes arising from this study have commenced generating peer reviewed research publications.
Summarising the main results, the design process is driven by geometry generation, calculation of heat transfer & flow, structural analysis, vibration and fatigue testing; the ESR team worked towards an integration of these for the benefit of the lifetime of heat exchangers. The practical lesson learnt is that there is a tendency for technical specialists to “work in isolation” until brought together by the design constraints. This highlights the industrial need to have specialist technology roles integrated within design teams possessing good communication.
ESR1 focused on the development of novel multi-disciplinary design optimisation methods, combining together a digital twin design and monitoring process. Automation of geometry generation and the multi-disciplinary analysis of heat exchangers, integrated the lessons learned from the other ESRs on fatigue and dynamic response prediction, to enable more efficient and robust heat exchangers to be designed more rapidly.
The work of ESRs 2 and 3 provided knowledge on the dynamic characteristics of heat exchangers. This allows for more accurate estimation of vibrational properties of these structures during design and in operation. A hierarchical design philosophy relying on simple but approximate models at early stages, followed by more elaborate but computationally expensive models is the way forward within a typical design cycle. These multi-fidelity models would improve the durability of these components by ensuring critical frequencies of the engine operation are avoided, given fixed computational resources.
Variations in the process parameters during the additive manufacture of stainless steel components have been found by ESR4 to affect defect spatial and size distributions. This was linked to the control of the relative effects of initiation, growth and coalescence processes in determining fatigue lifetimes. Future steps include the assessment of such features on microstructure and defect distributions in additively manufactured (AM) structures to aid the structural design assessment and optimisation of more complex designs.
Surrogate models provide an interesting and inexpensive way of providing information on the performance of an engineering system on which optimal designs, e.g. using genetic algorithms, could be searched. The framework developed by ESR1 could be used for a single or multiple objectives as well as multiple domains of physics (e.g. heat transfer, flow calculations, stress analysis, vibration and fatigue prediction). A combination of work from ESR2 and ESR4 has provided a method to feed vibration calculations of a structure (theoretically obtained or experimentally measured) to fatigue calculations, integrating structural dynamics and material behaviour. The knowledge generated by the application and development of experimental methods to test industrial components under ambient excitation by ESR3 should also be applicable to other industries with potential use for continuous condition monitoring and digital twin modelling.
The work of ESRs 2 and 3 provided knowledge on the dynamic characteristics of heat exchangers. This allows for more accurate estimation of vibrational properties of these structures during design and in operation. A hierarchical design philosophy relying on simple but approximate models at early stages, followed by more elaborate but computationally expensive models is the way forward within a typical design cycle. These multi-fidelity models would improve the durability of these components by ensuring critical frequencies of the engine operation are avoided, given fixed computational resources.
Variations in the process parameters during the additive manufacture of stainless steel components have been found by ESR4 to affect defect spatial and size distributions. This was linked to the control of the relative effects of initiation, growth and coalescence processes in determining fatigue lifetimes. Future steps include the assessment of such features on microstructure and defect distributions in additively manufactured (AM) structures to aid the structural design assessment and optimisation of more complex designs.
Surrogate models provide an interesting and inexpensive way of providing information on the performance of an engineering system on which optimal designs, e.g. using genetic algorithms, could be searched. The framework developed by ESR1 could be used for a single or multiple objectives as well as multiple domains of physics (e.g. heat transfer, flow calculations, stress analysis, vibration and fatigue prediction). A combination of work from ESR2 and ESR4 has provided a method to feed vibration calculations of a structure (theoretically obtained or experimentally measured) to fatigue calculations, integrating structural dynamics and material behaviour. The knowledge generated by the application and development of experimental methods to test industrial components under ambient excitation by ESR3 should also be applicable to other industries with potential use for continuous condition monitoring and digital twin modelling.