Exploration of predictive and Opportunistic Maintenance for measurement equipment

 Overview

Company Description 

ASML is the global leader in photolithography machines used to produce semiconductor chips. Its systems, especially those using extreme ultraviolet (EUV) light, enable the miniaturization of modern chips and drive the development of AI. Each machine integrates precision optics, mechatronics, software, and control systems at nanometer accuracy. ASML exemplifies challenges and state-of-the-art in managing highly complex systems and global supply chains, with thousands of specialized suppliers contributing critical components.

Project Description 

The reflectivity performance of EUV Source Collector Mirror, a critical component in the EUV Source, needs to be measured by a complex equipment called CEMT.

Semiconductor manufacturing demands high equipment reliability and operational precision, where any disruption can carry significant cost implications. The CEMT system, as a key component in baseline qualification workflows, requires ongoing attention to maintenance and part replacement to ensure consistent performance.
With the planned deployment of four additional CEMT systems across global sites, both maintenance requirements and logistical coordination are expected to increase. This added complexity may influence repair timelines and resource allocation, potentially affecting overall operational efficiency.

Goals of the Project  

This project aims to enhance the availability of CEMT systems, reduce spare part inventory costs, and prevent unplanned downtime through predictive and opportunistic maintenance strategies. The first step will be identifying module/parts suitable for predictive maintenance. Machine data will be analyzed to characterize degradation and failure patterns, and maintenance logs will be explored to understand maintenance intervals and spare parts demand. Then, a predictive model of the module/parts will be developed to anticipate its degradation. Hybrid models blending system knowledge and historical data will be used. Finally, an opportunistic predictive maintenance strategy will be proposed considering system health and performance, spare part inventory, lead time, operational schedule, and the predictive capability.

Predictive maintenance aims to use predictive models of machine failure/degradation to optimize maintenance schedules, inventory management, and resource allocation. A straightforward example is to use remaining-useful-life estimation to decide when to replace a component or when to order a spare part.

Opportunistic maintenance aims to optimize maintenance schedules considering system operations, scheduled/unscheduled downtime caused by other components, in order to reduce overall downtime and cost. A typical example is to align scheduled downtimes with the usage schedule of the CEMT system, avoiding CEMT from being down when it is needed.

Deliverables

• Module/part candidates suitable for opportunistic predictive maintenance
• Analysis results of machine data and maintenance log
• A trained predictive model
• Opportunistic predictive maintenance framework
• ASML internal report
• TUe master thesis (public)

Essential Student Knowledge

• Self-motivated, curiosity-driven attitude to lead the project
• Knowledge of maintenance processes and spare part management
• Knowledge of reliability engineering (e.g., Weibull, six-sigma, hypothesis-test, regression, etc.)
• Skills in data processing and model building
• Understanding and knowledge of formulating and solving optimization problems
• Basic understanding of physics and curiosity about highly complex technical systems

More information: escf@tue.nl