Fraunhofer Institute of Optronics, System Technologies and Image Exploitation IOSB 
Short description of the project
Production processes, especially in small and medium-sized enterprises (SMEs), must meet ever increasing demands on productivity, flexibility and quality of the manufactured products. However, the increasing complexity of the processes means that plants are often not optimally operated in terms of their efficiency. This results in an increased reject rate, high energy costs or reduced product quality. Particularly in SMEs, the plant efficiency achieved is highly dependent on the expert knowledge of experienced machine setters. Systematic process optimization, which includes data management and the evaluation of historical process data, for example, is often not possible due to the personnel infrastructure. The same applies to condition monitoring of the process. The plant monitoring carried out is often only very rudimentary, e.g. by selecting heuristic threshold values. The selection of threshold values is again based on the knowledge of the machine setter. Depending on their knowledge, disturbances in the process are detected early or too late.
This is exactly where the developed analysis platform comes into play. By means of the developed and integrated data management and especially thanks to the advanced data evaluation, it is possible to carry out a systematic data evaluation as well as plant monitoring and thereby achieve an increase in efficiency of the production process.
Simple connection of the platform to the company's IT infrastructure
The platform has standardized interfaces for connecting to different production processes. A complex installation process for maintenance and commissioning is not necessary. As an interface, the focus is primarily on the OPC UA standard. However, further interfaces can also be added to the platform on request.
Project goals
Real-time capable cross-plant monitoring and optimization
In comparison to traditional desktop applications, where evaluation and display are performed on the same computer, the platform has a modern client-server architecture. The production process can thus be monitored and historical data evaluated regardless of location and in real time.
Multi-user capability of the user interface
On the provided surface several users can use the platform at the same time. If a client communicates a user input to the server (e.g. acknowledging a displayed alarm status), this is forwarded to the other clients. At the same time, a new client queries the current information of the other clients when logging on and thus contains all relevant, up-to-date plant information.
Machine learning methods especially for SMEs
Machine learning methods usually have a large number of free variables. A parameterization of these variables can quickly become time and cost intensive. At the same time, parameterization requires a high degree of expert knowledge. The procedures provided in this platform use the smallest possible number of freely parameterizable variables. Most of the required variables are automated and estimated based on current and historical process records.
Project result
Development and testing of the platform using modern Industrie 4.0 equipment
The demo system for unit load production (see picture) available at the IOSB has all the requirements of a modern Industrie 4.0 system and is used for developing and testing the platform. In contrast to a classical system with one control unit, the demo system has one control unit per component (10 components in total), which are networked with each other. The data communication is done via OPC-UA. To be able to connect existing components that do not yet implement an OPC-UA interface, OPC-UA wrappers are available. This gives you the possibility to test a retrofit of old systems to the current state of the art.