A rapid and effective method for alloy materials design via sample data transfer machine learning

Abstract

One of the challenges in material design is to rapidly develop new materials or improve the performance of materials by utilizing the data and knowledge of existing materials. Here, a rapid and effective method of alloy material design via data transfer learning is proposed to efficiently design new alloys using existing data. A new type of aluminum alloy (E2 alloy) with ultra strength and high toughness previously developed by the authors is used as an example. An optimal three-stage solution-aging treatment process (T66R) was efficiently designed transferring 1053 pieces of process-property relationship data of existing AA7xxx commercial aluminum alloys. It realizes the substantial improvement of strength and plasticity of E2 alloy simultaneously, which is of great significance for lightweight of high-end equipment. Meanwhile, the microstructure analysis clarifies the mechanism of alloy performance improvement. This study shows that transferring the existing alloy data is an effective method to design new alloys.

Maximizing Triboelectric Nanogenerators by Physics‐Informed AI Inverse Design

“…Although conductive materials can be optimized by mixtures, there are major difficulties in obtaining the optimal triboelectric polarity and surface charge density due to the uncertainty of the key factors in compositions. [34,35] In addition, several technical challenges still need to be addressed to expand the implementation of TENGs. These challenges include ensuring their biocompatibility, enabling their operation in extreme conditions (such as high temperature, humidity, moisture, and varying excitation amplitudes and frequencies), and addressing the lack of sufficient experimental data in these domains.…”

“…However, intuitive guideline is still insufficient to achieve desirable electrical response as the complex nature of TENGs results in the difficulty of characterizing all the factors and accurately obtaining the relationships only using traditionally statistical tools. Addressing these technical challenges, AI-based performance prediction and optimization have recently been reported to tailoring the micro-compositions of functional materials to obtain desirably electrical response, [34] and unveil the relationships between the inputs and outputs without quantitatively modeling. [35,36] Table 2 summarizes the existing AI models developed for the performance prediction and output data analysis of four modes of TENGs.…”

“…AI tools have recently been applied to design TENGs with optimal electrical performance, for example, electro-physiological sensors, [68] TENG arrays, [69] self-powered sensor networks, [70] etc. However, [34][35][36], 2023, Wiley). [34][35][36] the large amount of data required for calibrating the AI models has severely limited their development.…”

“…However, [34][35][36], 2023, Wiley). [34][35][36] the large amount of data required for calibrating the AI models has severely limited their development. This is particularly a challenge for the AI algorithms that significantly rely on data such as NN and DL.…”

“…The dearth of adequate data at every stage of developing resilient AI models for TENG-based systems continues to be a pressing concern. b) Uncertainty in characterizing the key factors in composition and the lack of quantitative relationships by traditional statistical tools, and the difficulty in optimizing conductive and dielectric materials by composition mixtures while controlling structural response by design of contact interfaces (Reproduced with permission[34][35][36], 2023, Wiley) [34][35][36].…”