Description

Problem

A semiconductor manufacturer produces microchips for customers in various industries, including automotive, manufacturing, computers, mobile phones and consumer electronics. Microchips are manufactured in clean rooms under strict supervision in more than 1,000 process steps to avoid failures or quality variations. Pumps for ultra-pure water preparation are particularly important for semiconductor production, but their defects were previously unpredictable and could lead to disruptions in production.

Solution

On-site acoustic monitoring was developed for pumps based on a standard edge gateway. The acquired data was processed by machine learning algorithms, transferred to a central data collection in the cloud and summarised in dashboards for clear presentation. The data-based, self-learning IoT solution provides digital monitoring of the pumps on clear dashboards. The workload of specialist personnel is reduced and, thanks to machine learning, an early and accurate determination of maintenance requirements is made possible. Production-critical key figures can be better adhered to, as the current operating status is known and fail-safety is increased.