High entropy alloy design for dissolvable tools using machine learning

可溶合金是在特定操作条件下可溶解的特殊金属合金的新兴领域。这些合金广泛应用于石油和天然气、生物医学和空间应用。该提案旨在使用机器学习（ML）方法来定义用于可溶解工具的新合金系统。这些工具必须具有独特的强度和理想的腐蚀速率。迄今为止，铝合金和镁合金已用于此类应用。高熵合金（HEAs）和近化学计量HEAs（这里都称为HEAs）是新一代材料，有望提供这种性能的组合。考虑到调整HEAs结构和组成的巨大可能性，本提案旨在开发一种机器学习模型，该模型将公共领域和专利的结果与这些可溶合金的期望应用目标相结合。这些知识将与热力学数据库工具相结合，以确定HEAs的新合金成分。实验工作和机器学习建模将交互进行。然后将使用阿尔伯塔大学开发的混合熔模铸造增材制造方法进行实验验证。随后将评估机械和腐蚀性能，以验证ML模型。在获得足够的数据后，将建立并测试用于预测HEAs机械和腐蚀性能的ML模型。该项目的可交付成果将是可溶解材料的新材料，基于机器学习的工具和基于特定材料要求的定向材料空间研究方法。

Dissolvable Alloys is an emerging area of special metallic alloys that dissolve under specific operating conditions. These alloys have found wide spread use in oil and gas, biomedical, and space applications. This proposal aims to use Machine Learning (ML) approaches to define new alloy systems for application of dissolvable tools. These tools must provide a unique combination of strength and desirable corrosion rates. To date, aluminum alloys and magnesium alloys have been used for such applications. High Entropy alloys (HEAs) and near stoichiometric HEAs (all referred to here as HEAs) are a new generation of materials that show promise to provide such a combination of properties. Given the enormous possibilities in tunning the structures and compositions of HEAs, this proposal aims to develop a machine learning model that will combine the results in the public domain and patents with the desired application goals for these dissolvable alloys. Such knowledge will be combined with thermodynamic database tools to identify new alloy compositions for HEAs. Experimental work and ML modelling will be interactively carried out. An experimental validation will then be undertaken using a hybrid investment casting additive manufacturing approach developed at the University of Alberta. Mechanical and corrosion properties will subsequently be evaluated to validate the ML model. With sufficient data acquired, ML models for predicting the mechanical and corrosion properties of HEAs will be set up and tested. The deliverables of the project will be new materials for dissolvable materials, machine learning based tools and methods for directed materials space research based on specific material requirements.