top of page


Evaluation of production processes in the context of Industry 4.0

The task

Sensor data are collected in the factory and are sent in real time from the machines to an OPC UA server. OPC UA is a machine to machine communication protocol that was created for industrial automation and process control. The sensor data should then be made available for integration with further data and efficient analysis.


The challenge

Data production takes place at high speed; a failure of the infrastructure for data provision quickly leads to data loss. In addition, large unstructured amounts of data accumulate after only a short time, which must be efficiently combined with structured inventory information and analyzed.


our solution

In the course of this project, an OPC UA client was developed that connects to an OPC UA server and taps sensor data. The sensor data are published in a data lake and stored in a measurement data scheme for complex analyzes such as root cause and predictive analytics. In addition, the sensor data is persisted in a database to enable SQL-like queries for business intelligence. Subsequently, the stored data are made accessible in order to map complex issues on a simple and scalable analysis of the sensor data.


The customer benefit

New data is automatically provided in different views and can be easily connected to other data sources. The customer receives a programmatic view of the machine data for the analysis and optimization of the factory processes.



Our role

  • Data engineering

  • Data science

  • Software development

Our activities

  • Connection of the OPC UA server through client development

  • Integration of client in NiFi, streaming of machine data

  • Persistence of machine data on Hive and HDFS (in ASAM ODS format)

  • Data exploration with the DSL in DaSense

Technologies & methods

  • Applications: Hadoop, DaSense, NiFi, Prosys, UAExpert

  • Databases: Hive-on-Tez LLAP

  • Languages / frameworks: Python, Shell, SQL, Spark, OPCUA

  • Methods: time series analysis

bottom of page