The demonstrator develops a Failure Mode and Effects Analysis (FMEA) Assistant tool to support the engineers to analyse the existing information efficiently. FMEA Assistant is created by using existing data from the manufacturing process like structured or semi-structured FMEA, Failure Analysis (FA), 8D documents, and other domain-specific unstructured texts like production tools manuals, handbooks, and process descriptions patents, and similar. In FMEA Assistant, the available documents are AI engineered to address the consistency issues and incompleteness of available documents and generate the common view on single risks of different process steps in the semiconductor manufacturing process.
Beyond state-of-the-art developments and impacts
Current state-of-the-art solutions for risk management and root cause analysis in semiconductor manufacturing rely on human experts. Due to the increasing complexity and amount of information, this concept contains the risk of incompleteness. With the FMEA Assistant, the knowledge, which is currently distributed over many text documents produced by the individual human experts, is collected, organised, and accessible to users who are not necessarily experts in the field. This allows fast and reliable knowledge usage for risk assessment and troubleshooting in many different chip and process development cases in semiconductor manufacturing. The FMEA Assistant automates the root cause analysis task by using AI-based methods. This research is highly focused on the semiconductor industry FMEA analysis, and it is, according to our knowledge, the first of its kind.