Data flow and interfaces of the Process Information Management System. © (Image: RHI Magnesita)

In 2006, RHI decided to use a Process Information Management System, or PIMS in technical jargon, to evaluate and utilize its process data. The specific reason for this was the construction of a new plant in the Chinese city of Dalian, the commissioning of which was to be optimized from Austria in order to avoid having to send several teams of experts there for long periods of time. Following the successful initial deployment at the plant in Hochfilzen, the solution was to be implemented in Dalian and later rolled out globally.

Production parameters tracked and understood
The most important task of the PIMS was defined as tracking, analyzing and understanding production parameters such as electricity and gas consumption or pressing pressures in the very heterogeneous machine landscape live and historically. The system was to support production as well as commercial decision-makers. It also needed to archive process data fully automatically and provide both real-time and historical data. Ease of use and the option of graphical processing, calculations and evaluations were further criteria. Access via RHI Magnesita's global WAN (Wide Area Network) based on MPLS with a prioritized network connection was important for data exchange.

With its PI system, OSIsoft fulfilled the tender criteria perfectly. It is able to connect the control data via the OPC standard used in automation technology with the SAP modules, production data acquisition and evaluation databases as well as office systems. The system structures the data in such a way that engineers and managers on site can use the insights gained to increase productivity or use them in applications such as predictive maintenance to accelerate deep learning and the analysis of machines.

The central elements are the PI server in Vienna and the respective PI interface, which is installed at each production site and transmits the process data to the PI server in Vienna via TCP/IP.

Ready for operation in a week
After the solution had proven its performance in initial tests in Vienna and Hochfilzen, the development partner and RHI's information management team also installed it in Dalian within just one week. Simple, redundant miniservers are used as PI interfaces. The hot standby technology guarantees that if one system fails, the system immediately switches over to the parallel computer. The company relies on a virtual server landscape that has been adapted to its growing requirements over the years - the server has grown along with the global rollout.

The PI system collects data, connects the individual systems and analyzes them. © (Image: OSIsoft)

In order to be able to read out all control data correctly, all machine controls had to be connected to the PI interface with their IP addresses and the so-called tags for all conceivable signal data had to be configured on the PI server.

Around 430 machines in 26 production plants are currently integrated into the PIMS, sending around 70,000 data signals. "Our experts can view the evaluations from anywhere. They can see firing and mixing temperatures, the mixer speed, the pressing pressure or the pressing time in real time and over time. In addition to improving energy efficiency, this saves travel time and immense costs," says Daniel Neubauer, Team Manager Business Applications R&D/Quality Management and PIMS application manager in the Information Management team. "A special feature of the OSIsoft solution is the intelligent storage technology, thanks to which not even a terabyte of data volume has been created even after more than ten years of use," adds Neubauer.

The stored history forms the basis for analyses of how the individual production parameters influence product quality and how they can be optimized. "Another issue is preventive maintenance. Previously, we serviced the machines at fixed intervals. Based on the counter readings available in PIMS, we can use performance-oriented maintenance by transferring the data to SAP on a daily basis and now carrying out maintenance every 100,000 press strokes, for example. This means that the machines are never serviced too early or - even worse - too late," says Neubauer. RHI Magnesita goes one step further with predictive maintenance, which uses process data to anticipate and report faults before a machine breaks down. By replacing a simple spare part in good time, operations can continue without interruption.

The recorded and evaluated information has also become an indispensable part of the continuous global optimization process in quality assurance and in the R&D area. For example, the system uses the raw data to calculate peak and hourly averages for gas and electricity consumption or deviations from specifications and makes them available to users on visualization monitors in production, from machine operators to group and plant managers. In an emergency, the process data also triggers alarm e-mails.

"Based on this data, we can not only quickly identify problems such as deviations in the process, but also subsequently eliminate them through simulations and tests. New developments are now also easier to implement across all areas of the company, because the data quality and availability gained enables knowledge to be exchanged virtually at the touch of a button," says Reiterer.