Framework for Identifying Failure Mechanisms in Consumer Electronics Using Large Language Models

As electronic waste becomes a growing concern, extending the lifetime of devices by increasing their durability has become a key lever for promoting sustainability. Although a durability index was proposed in a previous study, determining tests to cover the entire lifespan of target devices remains a challenge. To identify the necessary tests, it is essential to first identify potential failure mechanisms. This study proposes a framework for identifying failure mechanisms from brief information, such as the product description of the target device. The proposed framework leverages large language models (LLMs) to generate use cases and identify potential stresses and failure mechanisms. A modularized design is adopted for better flexibility in accommodating additional information between the modules.