For many, Smart Manufacturing is how ‘data analytics and artificial intelligence technology are used to increase efficiency and automation’. However, this definition will not drive the best outcome in the vacuum and abatement system because it misses an important component: the human side of smart manufacturing, specifically the knowledge and experience required to apply the necessary context.
So perhaps better expressed, Smart Manufacturing is the application of data analytics and technology to connect people, products, and processes to drive the best results.
To achieve this, we provide 3 key capabilities:
We provide 24/7 data monitoring and collection from both the SubFab and the clean room, for a wide variety of equipment makes and models, including almost all of Edwards equipment and a selection of non-Edwards equipment.
We have the capability to integrate with each customer’s key operating systems, to allow for effective information exchange. This requires the connection of not only the SubFab and clean room equipment, but also to other monitoring and process control software.
We use real-time and historical data analysis and data export tools, to drive context on system performance and improvements.
We analyse data to transform it into actionable insights. This involves understanding why down events happen and what can be changed to achieve the best outcome. This requires in-depth knowledge of process chemistries and vacuum physics, two key components of our domain knowledge.
We then implement actionable insights with models, such as best-known-methods or standard operating procedures. We assist both engineers and management teams to make informed decisions and provide reliable reporting data on key performance indicators.
The objective is to drive continuous equipment and process performance improvements across the entire Fab.
We developed predictive maintenance methods to reduce or eliminate unscheduled down events or catastrophic failures, which result in costly wafer loss. Preventative maintenance regimes have to be conservative, which may result in unnecessary maintenance and extra cost in maintaining equipment that might not be about to fail. Run-to-fail maintenance is also costly because it increases the costs to repair and, on failing, product may be lost.
The optimal method is a shift to a predictive maintenance programme that will removes unnecessary costs, reduces the risk of wafer loss and optimises uptime and Fab performance.
Predictive maintenance is brought to life with robust and precise models. These include ‘Remaining Useful Life’ predictions and machine learning methods, that issue maintenance guidance. We may recommend actions in various alert states which range from ‘advisory’, ‘warning’ to ‘high risk of failure’ warnings.
Our predictive maintenance models grant customers the flexibility to synchronise maintenance across their entire Fab maximising production and preserving the total cost of ownership of the facility’s operations.