Industrial Data Pipelines for Manufacturing Applications

Abstract

Small and medium enterprises (SMEs) struggle with the uptake of industrial data in their business. SMEs have expressed interests to integrate physical legacy machines in factories and uptake Industrial Internet of Things (IIoT) for data-driven industrial processes. A widely recognizable aids for companies are testbeds or test-before-invest environments where companies can familiarize themselves with Industry 4.0 (I4.0) technologies. In this paper, an easily scalable approach to collect industrial data is presented. The environment can be used by SMEs to benchmark the developed solutions using affordable licenses and test aspects of the industrial data best suitable for their purposes before making investments. In TAMK FieldLab I4.0 pilot trial, horizontal integration is done in between 14.0 field and control layer to integrate different machines' data to common database and common format to be used for visualization and later for data analysis. The focus of the paper is on data connectivity capabilities. The trial utilize data from different machines. First phase includes exporting data from different machines to same data table with time stamps, to enable comparison of data with same time base. With common time base it's possible to produce synchronized timeseries data visualization from all TAMK FieldLab's machines. Used common database is PostgreSQL, where data from all machines is stored. For horizontal field level integration minimum amount APIs and protocols is used. For common API protocol MQTT was chosen as it's lightweight and supported widely. ABB robot data is taken trough OPC/UA and MQTT. From Omron mobile robot data is transferred trough ARCL. The paper will identify gap areas of the implementation in comparison with I4.0 and when appropriate, RAM 4.0 framework. Finally, the next steps that will be taken in near future regarding artificial intelligence (AI) and upgrading the IIoT system to match with RAMI 4.0 framework will be presented.