Predicting FEA Results of Cable Trays Using Machine Learning

In the design process of cable trays in nuclear power plants, mechanical analysis is a core technology used to ensure the safety and reliability of structural design. Traditional mechanical analysis methods, especially finite element analysis, although accurate, often incur high computational costs and time consumption. With the development of machine learning (ML) technology, it has shown potential in various engineering applications to effectively improve analysis efficiency and optimize designs. Particularly in addressing complex nuclear engineering problems, ML can learn from large datasets to predict system behavior under unknown conditions, thereby assisting in decision-making and design optimization. This paper focuses on a limited dataset of nuclear engineering cable trays. It first designs data augmentation algorithms to establish logical relationships between cable tray data, then proceeds with normalization and sample clustering. Multiple ML algorithms are experimented with, demonstrating effective prediction of output parameters of cable tray models within a certain margin of error.