By susanne.gutermuth@cosmino.de In
Beside the traditional evaluation of historical data using indicators and reports, MES customers of COSMINO AG in Nuremberg increasingly require advanced and predictive analytics. In addition to the Business Intelligence module of the Cosmino MES Plus solution, this requires technologies, which enable Big Data analysis and the detection of hidden correlations and dependencies.
COSMINO AG uses time series databases for such tasks, which optimize efficient storage and fast access. The first prototype of the Cosmino Big Data solution is based on OpenTSDB (free Apache license) and manages the data in a Hadoop Data Lake (free Apache license). This replaces expensive licenses and long loading times, which is an aspect not to be underestimated for mass data, such as process parameters and energy consumptions.
Thanks to this technology, the Cosmino Big Data solution is ready for machine learning algorithms. The development of proprietary, customer-specific algorithms is planned with users within the scope of joint use cases. Application cases are predictive maintenance, smart manufacturing, and anomaly detection. Corrective measures resulting from the Big Data analysis can be implemented via Cosmino MES Plus, for example, by adjusting the maintenance schedule automatically.
