A new method using reinforcement learning for designing bioinspired composite materials is proposed. While bioinspired design of materials is a promising avenue, the possible combination of building blocks in a composite is usually intractable. In this work, a new method is proposed based on reinforcement learning applied as an autonomous agent for arranging the microstructure that is composed of brittle and soft material. The resolution of the design space is enhanced in a progressive fashion, reaching increasingly higher resolution. The results show that the resulting high-resolution designs can significantly reduce stress concentrations at crack tips and enhance mechanical resilience. Complementing the experimental results with manufactured optimal composites shows excellent agreement with the optimal results obtained using the AI method. The framework reported in this work may serve as an alternative to conventional composite optimization techniques, which often suffer from the curse of high dimensionality and are also unable to effectively predict high-resolution designs, due to limitations of the algorithms to escape low-resolution local minima. The new approach discussed in this work can be widely applied in multiple areas of engineering and design, and the progressive multiresolution approach may also be critical for the de novo autonomous reinforcement engineering solutions.

The demand for highly specific and complex materials has made the development of controllable manufacturing processes crucial. Among the numerous manufacturing methods, casting is important because it is economical and highly flexible regarding the geometry of manufactured parts. Since solidification is an important stage in the casting process that influences the properties of the final product, the development of a controllable solidification process using modeling methods is necessary to create superior structural properties. However, traditional modeling methods are computationally expensive and require sophisticated mathematical schemes. Therefore, a deep learning model is proposed to predict the morphology of the dendritic crystal growth solidification process, along with a reinforcement learning model to control the solidification process. By training the deep learning model with data generated using the phase field method, the solidification process can be successfully predicted. The crystal growth structures are designed to be altered by adjusting the degree of supercooling in the deep learning model while implementing reinforcement learning to control the dendritic arteries. This research opens new avenues for applying artificial intelligence to the optimization of casting processes, with the potential to utilize it in the processing of advanced materials and to improve the target properties of material design.

No abstract