At the Laboratory for Machine Tools and Production Engineering (WZL) of RWTH Aachen University, scientists, mathematicians, and software developers conduct manufacturing research, working together to gain new insights from machine, product, and manufacturing data. Manufacturers partner with the team at WZL to refine solutions before putting them into production in their own factories.
Recently, WZL has been looking for ways to help manufacturers analyze changes in processes, monitor output and process quality, then adjust in real-time. Processing data at the point of inception, or the edge, would allow them to modify processes as required while managing large data volumes and IT infrastructure at scale.
Connected devices generate huge volumes of data
According to IDC, the amount of digital data worldwide will grow by 23% through 2025, driven in large part by the rising number of connected devices. Juniper Research found that the total number of IoT connections will reach 83 billion by 2024. This represents a projected 130% growth rate from 35 billion connections in 2020.
WZL is no stranger to this rise in data volume. As part of their manufacturing processes, fine blanking incubators generate massive amounts of data that must first be recorded at the sharp end and processed extremely quickly. Their specialized sensors for vibrations, acoustics and other manufacturing conditions can generate more than 1 million data points per second.
Traditionally, WZL’s engineers have processed small batches of this data in the data center. But this method could take days to weeks to gain insights. They wanted a solution that would enable them to implement and use extremely low-latency streaming models to garner insights in real-time without much in-house development.
Data-driven automation at the edge
WZL implemented a platform which could ingest, store, and analyze their continuously streaming data as it was created. This system gives organizations access to a single solution for all their data (whether streaming or not) that provides out-of-the box functionality and support for high-speed data ingestion with an open-source and auto-scaling streaming storage solution.
Now, up to 1,000 characteristic values are recorded every 0.4 milliseconds – nearly 80TB of data every 24 hours. This data is immediately stored and pre-analyzed in real-time at the edge on powerful compact servers, enabling further evaluation using artificial intelligence and machine learning. These characteristic values leverage huge amounts of streaming image, X-ray and IoT data to detect and predict abnormalities throughout the metal stamping process.
The WZL team found that once the system was implemented, it could be scaled without constraint. “No matter how many sensors we use, once we set up the analytics pipeline and the data streams, we don’t have to address any load-balancing issues,” said Philipp Niemietz, Head of Digital Technologies at WZL.
With conditions like speed and temperature under constant AI supervision, the machinery is now able to automatically adjust itself to prevent any interruptions. By monitoring the machines in this way, WZL have also enhanced their predictive maintenance capabilities. Learn more about how you can leverage Dell Technologies edge solutions.
Intel® Technologies Move Analytics Forward
Data analytics is the key to unlocking the most value you can extract from data across your organization. To create a productive, cost-effective analytics strategy that gets results, you need high performance hardware that’s optimized to work with the software you use.
Modern data analytics spans a range of technologies, from dedicated analytics platforms and databases to deep learning and artificial intelligence (AI). Just starting out with analytics? Ready to evolve your analytics strategy or improve your data quality? There’s always room to grow, and Intel is ready to help. With a deep ecosystem of analytics technologies and partners, Intel accelerates the efforts of data scientists, analysts, and developers in every industry. Find out more about Intel advanced analytics.