Using process mining in combination with (more regular) data analysis to predict the chance of part failure during EUV assembly and testing.
Introduction
ASML is a worldwide market leader in the production of lithography systems. These systems are used by customers like Intel and Samsung for their production of integrated circuits. Lithography systems are large (container sized), expensive (between 15 and 100 million dollar) and high-tech (over 50% of 16k FTE work at Development & Engineering). Introduction of new system types has been a key competitive driver of ASMLs growth over the last 30 years. The EUV Factory produces ASMLs latest system types. You can find more information about ASML here and here.
Assignment
The EUV factory experiences part related quality issues during the assembly and qualification of our systems. There are projects ongoing to reduce these quality issues by finding them as early as possible in the supply chain. Reasons for these quality issues (broken parts) are diverse. They range from handling problems in warehouses to wrong packaging defined in the part administration. Another potential problem is the usage of refurbished (repaired, formerly broken) parts. What all these problems have in common is that they might be found beforehand by combining part characteristics as captured in regular part documentation (SAP database) and knowledge about the history of where the part has been before in enters the factory. We believe we can use process mining in combination with (statistical) analyses to find some of these quality issues before they cause problems in the factory. It would be interesting for us to know to which extend this is possible and how we could go about doing so in a more structural way in the future.