PRIME – Process Refinement for Improved Manufacturing Efficiency
We aim to develop a sustainable, scalable framework to boost production performance, enable multiline integration at ABB, and deploy an LLM-based chatbot for intuitive interaction with the solution. All are driving toward zero-defect manufacturing and reduced scrap rates required for green production.
Description
In today’s manufacturing environments, achieving Zero-Defect Manufacturing (ZDM) remains challenging due to ambiguous quality labels, variable batch structures, and limited lifecycle management of AI models. ABB’s production lines exemplify these issues, where inconsistent data and isolated AI systems restrict scalability and reliability.
The PRIME project aims to overcome these barriers by developing a sustainable, adaptive, and standardized AI framework that links all stages of production. Its objectives are to boost production performance, enable multi-line integration, and move toward zero-defect and low-scrap manufacturing.
PRIME introduces adaptive AI capable of learning from uncertain and dynamic data, integrates human feedback through explainable interfaces and LLM-based chatbots, and embeds full lifecycle and interoperability management using the Asset Administration Shell (AAS). This approach ensures robust, traceable, and transferable AI deployment across ABB sites, paving the way for scalable, green, and efficient industrial production.
Key data
Projectlead
Marcin Sadurski, Stefan Waskow (ABB Schweiz AG)
Deputy Projectlead
Project team
Maryam Rezayati, Fatemeh Mohammadi Amin, Rafael Monteiro Marques, Patrick Haezi (ABB Schweiz AG), Marko Barutcu (ABB Schweiz AG), Afrooz Laghaei (ABB Schweiz AG)
Project partners
ABB Schweiz AG / ABB Schaffhausen
Project status
ongoing, started 11/2025
Institute/Centre
Institute of Mechatronic Systems (IMS)
Funding partner
Innosuisse Innovationsprojekt
Project budget
1'005'848 CHF
