In this groundbreaking study, advanced variants of the Deep Operator Network (DeepONet) architecture were utilized to predict entire thermal and mechanical solution fields in materials processing and additive manufacturing. These newly devised DeepONet formulations incorporated sequential learning and Residual U-Net (ResUNet) architectures. They were trained for the first time to simultaneously forecast complete thermal and mechanical solution fields under variable loading conditions, loading histories, process parameters, and varying geometries. The research showcased the efficacy of DeepONets in providing accurate solutions across different application areas - from steel continuous casting with multiple visco-plastic constitutive laws to sequentially coupled direct energy deposition for additive manufacturing. Despite the complexity of spatially variable stress distributions, DeepONets demonstrated the ability to infer full-field temperature and stress solutions rapidly and efficiently. They outperformed traditional finite-element analysis (FEA) methods by several orders of magnitude. Looking ahead, future studies will focus on incorporating data from every simulation time step into neural networks to predict the complete history of solution fields. Additionally, there are plans to explore additive manufacturing for intricate 3D structures in order to accurately predict mechanical properties. The utilization of high-performance computing (HPC) systems will be further optimized to generate data from multi-scale computational frameworks encompassing micro and mesoscale phenomena. The research findings underscored that well-trained advanced DeepONet architectures can serve as a surrogate for modeling complex real-world manufacturing processes. This breakthrough opens up avenues for preliminary optimization, inverse analysis, sensitivity analysis, uncertainty quantification, online controls, digital twins creation, and other computationally intensive numerical analyses requiring multiphysics forward solution evaluations in domains with variable geometries and input parameters. To facilitate replication of results and promote transparency in research practices,the data and source code supporting this study will be made available in a public GitHub repository. Furthermore, the authors have declared no conflict of interest in conducting this study.
- - Advanced variants of Deep Operator Network (DeepONet) architecture used for predicting thermal and mechanical solution fields in materials processing and additive manufacturing
- - Newly devised DeepONet formulations incorporated sequential learning and Residual U-Net (ResUNet) architectures
- - Trained to simultaneously forecast complete thermal and mechanical solution fields under variable conditions, loading histories, process parameters, and geometries
- - Demonstrated efficacy in providing accurate solutions across different application areas
- - Outperformed traditional finite-element analysis methods by several orders of magnitude
- - Future studies will focus on incorporating data from every simulation time step into neural networks and exploring additive manufacturing for 3D structures
- - Utilization of high-performance computing systems optimized for multi-scale computational frameworks
- - Well-trained DeepONet architectures can serve as a surrogate for modeling complex real-world manufacturing processes
- - Data and source code supporting the study will be made available in a public GitHub repository
Summary- Scientists created smart computer programs to help predict how materials change when they are heated or put under pressure.
- These programs learn as they go and use a special design called Residual U-Net to make accurate predictions.
- They can predict how materials will behave in different situations and have been proven to work well in many areas.
- These programs are much better than older methods at figuring out what will happen to materials.
- In the future, researchers want to make these programs even smarter by using more data and exploring new ways of making things.
Definitions1. Deep Operator Network (DeepONet) - A type of advanced computer program used for predicting changes in materials during manufacturing processes.
2. Formulations - The specific ways in which something is made or designed.
3. Sequential learning - Learning from data in a step-by-step manner.
4. Residual U-Net (ResUNet) - A specific architecture used in computer programs for making predictions based on previous information.
5. Finite-element analysis - A traditional method for predicting how materials will behave under different conditions.
6. Neural networks - Computer systems that learn and make decisions based on patterns and data.
7. Additive manufacturing - Creating objects by adding material layer by layer using technology like 3D printing.
8. High-performance computing systems - Powerful computers optimized for handling complex calculations quickly and efficiently.
9. Surrogate - Something that stands in for or represents something else, like a model representing a real-world process.
Introduction:
The field of materials processing and additive manufacturing has seen significant advancements in recent years, thanks to the integration of advanced technologies such as artificial intelligence (AI) and deep learning. In particular, the use of Deep Operator Networks (DeepONets) has shown promising results in predicting thermal and mechanical solution fields in these processes. This groundbreaking study aims to further enhance the capabilities of DeepONets by incorporating sequential learning and Residual U-Net architectures.
Background:
Materials processing and additive manufacturing are complex processes that involve various parameters such as loading conditions, loading histories, process parameters, and varying geometries. Traditional methods for predicting thermal and mechanical solution fields, such as finite-element analysis (FEA), can be time-consuming and computationally intensive. This is where DeepONets come into play – they have the ability to rapidly infer full-field temperature and stress solutions with high accuracy.
Methodology:
In this study, advanced variants of the DeepONet architecture were trained for the first time to simultaneously forecast complete thermal and mechanical solution fields under variable conditions. The newly devised formulations incorporated sequential learning techniques that allowed them to learn from previous simulations at different time steps. They also utilized Residual U-Net architectures which have been proven effective in image segmentation tasks.
Results:
The research showcased the efficacy of DeepONets in providing accurate solutions across different application areas – from steel continuous casting with multiple visco-plastic constitutive laws to sequentially coupled direct energy deposition for additive manufacturing. Despite the complexity of spatially variable stress distributions, DeepONets demonstrated superior performance compared to traditional FEA methods by several orders of magnitude.
Future Directions:
Looking ahead, future studies will focus on incorporating data from every simulation time step into neural networks to predict the complete history of solution fields. Additionally, there are plans to explore additive manufacturing for intricate 3D structures in order to accurately predict mechanical properties. The utilization of high-performance computing systems will also be further optimized to generate data from multi-scale computational frameworks encompassing micro and mesoscale phenomena.
Implications:
The research findings underscored that well-trained advanced DeepONet architectures can serve as a surrogate for modeling complex real-world manufacturing processes. This breakthrough opens up avenues for preliminary optimization, inverse analysis, sensitivity analysis, uncertainty quantification, online controls, digital twins creation, and other computationally intensive numerical analyses requiring multiphysics forward solution evaluations in domains with variable geometries and input parameters.
Data Availability:
To facilitate replication of results and promote transparency in research practices,the data and source code supporting this study will be made available in a public GitHub repository. This will allow other researchers to validate the findings and build upon them for future studies.
Conclusion:
In conclusion, this groundbreaking study has demonstrated the potential of DeepONets in predicting entire thermal and mechanical solution fields in materials processing and additive manufacturing. The incorporation of sequential learning techniques and Residual U-Net architectures has further enhanced their capabilities. With continued advancements in AI technology, it is expected that DeepONets will play a crucial role in optimizing these processes and improving overall efficiency.