Intelligent process monitoring in battery production
Festo further expanded its AI expertise with the acquisition of Resolto Informatik in April 2018. Together with Scraitec, Resolto has developed a software solution that analyzes and interprets data in real time and detects and reports anomalies. The system also continuously learns through ongoing data analysis and expands its knowledge base. This machine learning makes intelligent process monitoring possible.

Visitors to the Hannover Messe were able to experience the software solution live. The application shows the detection of faulty batteries. A handling portal lifts the batteries. In combination with the new CPX-E-CEC modular controller and the new CMMT-AS servo drive controller, monitoring is possible in real time. The Resolto monitoring software monitors the motor currents and position values of the axis. If anomalies occur, for example if the handling system picks up an incorrect battery format, a message is generated.

Data can be collected and monitored by the intelligent software solution either on edge or on premises, or via the IoT gateway CPX-IOT in the Festo Cloud. The use of AI on edge or on premises means that all data remains in-house, without security risks or delays in data streams due to network latency. It is important that sufficiently structured data is available in order to be able to carry out a meaningful analysis with the AI tool. The cloud, on the other hand, with its very high computing capacities, offers good evaluation results across multiple, distributed production sites.

The programming effort for process monitoring and error handling is significantly reduced thanks to AI. This provides the customer with valuable know-how in real time. Faulty parts and processes as well as machine failures can be detected and prevented at an early stage in the production process. Another advantage is the complete transparency and traceability of process anomalies to the respective piece produced. Large-scale recalls of entire series could become unnecessary in future, as the faulty part can be clearly identified and thus ejected in a targeted manner.

Intelligent component monitoring
The exhibit showed how a learning algorithm is used to monitor component faults. Data from an electrical axis is recorded and collected by the CMMT controller. The monitoring algorithm and the monitoring of the summarized data runs completely in the Festo Cloud. The data is therefore available anytime and anywhere.

Customers not only benefit from less programming effort. The normal states of their individual processes are taught into the learning algorithm during operation or from historical data. The data can be evaluated immediately via the cloud. This allows users to quickly identify deviations and trace the causes of errors directly. Replacement components can also be identified quickly. This saves time, reduces downtimes and lowers maintenance costs. In the future, it is conceivable that component groups such as modules, systems or entire machines could also be monitored in this way. as