With the advancing digitalization and the radically developing digital technologies within companies over the last decade, organizations in general had to think about new ways of working. Digitalization and the integration of (edge) AI plays an important role in searching for the best solution to a specific business need in various industries. Therefore, companies and in specific the semiconductor industry, is accelerating its digital transformation efforts related to factory automation and digitalization of core processes such as quality management processes.
The challenges of semiconductor manufacturing are manifold:
The production occurs continuously, 24 hours a day, 365 days a year, with more than 1000 structuring steps per wafer carried out using hundreds of different production tools (also known as equipment). These processes generate an enormous amount of production related data every day, providing a solid foundation for developing and applying innovative methods and tools that support automated decision-making to overcome the following challenges.
- Ever-increasing complexity in semiconductor manufacturing, involving several hundred individual process steps to produce a semiconductor chip.
- Teams with specific and defined technical domain knowledge related to their area of work.
- Large amount of production data generated.
- Efficient use of resources needed.
A digital tacit knowledge platform
The approach to develop and use an explainable AI model based on historic production quality data in form of a digital tacit knowledge platform to accelerate quality improvements in semiconductor manufacturing is pursued in the project demonstrator “Smart Semiconductor Manufacturing Assistant”, from Infineon. With this platform Infineon aims to:
- Fast identify and predict factors causing quality issues
- Identify and prioritize processes that improve quality
- List down potential improvement actions
- Use AI model trained and tested using historic production quality data
- Improve the interpretation of results to understand the factors influencing the measure of interest
- Choose the most practical improvement action
- Enabling real-time recommendations for engineers
The role of Infineon and project status
Within the EdgeAI project, Infineon Technologies Austria AG aims to generate a decision support tool to establish AI based causal links between different production databases (document data, production data and diagrams). This EdgeAI solution will be implemented in the demonstrator in Value Chain 1 “digital Industry”, to address the following:
- Increase the overall production efficiency
- guarantee high quality in production
- sustainability of the factory in a semi-autonomous production environment.
The vision of Infineon is to expand this innovative digital tacit knowledge platform to the FE-factory cluster in the context of Infineon’s one virtual fab concept to create a greater impact for Infineon and to contribute to a sustainable and resilient European semiconductor industry.
Blog signed by: IFAT team