The partners combine the digital twin platform 'sphinx open online' from GFT with edge computing from T-Systems and artificial intelligence from the cloud. According to the company, errors in production are detected up to 300% faster.

Increasingly large amounts of data are being generated in a smart factory. Manufacturers can adapt and optimize processes based on the analysis results of this data. With predictive maintenance, they can prevent expensive unplanned downtime in production. In order for companies to evaluate this data in real time, they usually need high network capacities between the production hall and the central data center. There they are analyzed and the results are fed back into production. The integrated solution from T-Systems and GFT based on edge computing and IIoT offers new fields of application for smart production.

In the joint solution from GFT and T-Systems, sensors record status data locally and in real time at the individual systems. The data is sent in parallel, but not time-critical, via smaller bandwidths to a central data center or via 5G networks to the cloud. Due to the significantly lower bandwidth, data transfer costs are reduced by up to 70%. The solution analyses the data collected at the edge and uses artificial intelligence (AI) to identify optimization potential.

"Previously, manufacturers had to combine several services, software and hardware into a complete solution themselves. Thanks to our partnership with GFT, they can now get the whole package from a single source," says Frank Strecker, responsible for T-Systems' public cloud business. "Edge computing from us, data collection, digital twin and AI analysis from GFT - this is how we can connect the production hall with the data center. We also offer our own public and private cloud services. Or the customer can opt for the public cloud of our partners AWS, Microsoft Azure or Google."

With 'sphinx open online', companies can monitor their systems and control them fully automatically. For example, they receive information about operating states such as temperatures or pressure, as well as resource consumption or current output. To do this, data from all relevant sources is transferred to a common model ("model in the middle"). In 'sphinx open online', digital twins are created from this. These virtual images control systems automatically. The digital twins can also be used to create far-reaching simulations. Users can modify entire systems and production processes and check the effects without having to physically touch a single machine. The simulations reveal optimization potential. These optimizations can be implemented automatically by comparing them with defined parameters and rules. For example, to adapt specific processes or to plan necessary maintenance windows more precisely. Manual intervention is no longer necessary, as all measures are fed back into the industrial plant and applied in real time.