Even in the most advanced semiconductor fabs today, the availability and uptime of SubFab and clean room vacuum systems are governed by run-to-crash and preventive maintenance strategies. They tend to be less efficient and cause unnecessary waste on energy, manpower, plus lower uptime and with extra spare parts consumed.
 
Edwards, world leader on vacuum technology, develops Predictive Maintenance (PdM) service model which tailors to the needs of sub fab to tackle the above mentioned issues through making tens of thousands of pumps and abatements in your fab intelligent. Edwards’ Predictive Maintenance strategies are empowered by Machine Learning (ML), a subfield of Artificial Intelligence (AI) and the keys to success are the combination of robust data and computer science.

Abatement is an essential equipment to reduce environmental emission in semiconductor fab. It requires regular field service resource to maintain. We developed a ML algorithm to predict equipment health which has been applied to 500+ maintenance events. Based on the algorithm & analysis, potentially 50% of the PMs were done unnecessary. AI models can help optimize maintenance effect to when it’s needed on the condition of the tools. This helps save time, resources on field service, energy and spare parts.

Another case to share is how PdM applied to dry pumps to optimize maintenance effect and avoid waste through unnecessary swaps on harsh CVD deposition process. With our ML algorithm applied, it suggests that only 35% of preventative pump swaps were shown to be timely, and 20% of the swaps happened after an unscheduled down could have been avoided and 45% of preventative swaps were shown to be too early.
 
Edwards is pioneering in vacuum systems ML applications by translating our extensive vacuum systems domain knowledge into mathematical models and prediction for maintenance guidance and prioritization. We continue to develop and deploy AI solutions to your subfab to help achieve fab operational excellence, lower cost, energy consumption and manpower to address the challenges of future human resource availability and increasing significance of ESG.