© Fotolia | zapp2photo



At the second VDMA "Predictive Maintenance" conference in Frankfurt am Main in February, 130 participants cast a glance into the future of maintenance in the context of Industrie 4.0.

Predictive maintenance keeps machines running. <br>© Fotolia | bobo1980In addition to production processes, Industrie 4.0 is increasingly shaping the way machines and components are maintained, and digitalization and interconnectivity also makes it possible to conduct predictive maintenance measures. The advantages of predictive maintenance are obvious - knowing exactly when and to what extent maintenance measures are required saves time and money. "Preventative maintenance aims at maximizing machine availability while at the same time minimizing costs as best possible," said Werner Binsmaier, Vice President Central Development at Homag Group AG, a manufacturer of woodworking machinery, at the VDMA's 2nd Predictive Maintenance 4.0 conference in Frankfurt am Main.

Until now, maintenance is mainly conducted at regular intervals, for example to replace a module that is expected to be nearing the end of its service life based on empirical values. Another option is to simply wait for a failure. "Sudden failures are the worst case scenario in large plants. Maintenance measures are thus aimed at preventing these things from happening in the first place," said Binsmaier.

Predictive maintenance will gain in importance in future, as a study conducted by the Roland Berger consulting firm clearly shows. Yet only one-quarter of the companies surveyed already has a strategy ready for implementation. Most companies stated that this was not because of technological aspects, but it rather comes down to the fact that they do not yet know how to define the added value.

Monitoring - predicting - preventing

The starting point of predictive maintenance is monitoring the status of a component using sensor technology. The data collected is then stored in a data cloud in the form of a digital twin and compared with the data predefined for the physical component in question. An alarm is raised if there are any deviations between these sets of data. "It all comes down to drawing the right conclusions based on the data," said Dr. Stefan Spindler, Member of the Executive Board of Schaeffler Technologies AG & Co. KG, an automotive and industrial supplier based in Herzogenaurach. Protecting data is another major challenge, he added, and there is still a need for standards for interfaces.

According to Dr. Roger Kehl, CIO of Festo AG & Co. KG, an automation specialist from Ostfildern-Scharnhausen, networking is the most important prerequisite for predictive maintenance. "The main objective is to define how a system is fed with data and how it can be retrieved. Predictions can only be made if all data is collected and stored in a cloud," he said. Many companies make use of internal data clouds, but shy away from external cloud service providers for fear of data misuse by third parties. "These concerns are largely unfounded because of the fact that data is stored in an unstructured manner in external clouds. Without any expert knowledge, third parties will hardly know what to make of this data," Kehl added.

External clouds offer the advantage that customers and suppliers have access to a larger data basis for making analyses and predictions. Kehl also believes that predictive maintenance might also become relevant in the after-sales business, as manufacturers will be able to charge the costs plants operators are saving thanks to predictive maintenance. "Predictive maintenance can thus be marketed and sold as an additional value," Kehl explained.

Need for action

Standards for interfaces still need to be developed. Without them, maintenance measures will be tied to specific suppliers and users will not have the option of selecting another supplier. Prof. Dr. Martin Wollschläger, who is researching this topic at the Institute of Applied Computer Science at TU Dresden, expects that agreement at European level is not likely to be reached before the end of 2018. "A global solution is rather unlikely for the time being, seeing as the USA and Japan are not willing to participate," he said. But it is very encouraging that VDMA has already forged ahead in standardizing predictive maintenance," the university professor said in praising the association's efforts.

Changed work environments

The digitalization of production processes will also introduce many changes to the employees' daily working life. In the case of predictive maintenance, this often leads to reduced workloads for the service personnel. "Our service staff no longer has to continuously monitor plants or aggregates. For example, they can use an app on their smartphones to check when and where maintenance measures need to be conducted," said Dr. Gunter Beitinger, Vice President Manufacturing at Siemens AG in Amberg.

In general, it is very important to keep employees on the path to digitalized plants motivated by continuously providing information, he added. "At our plant in Amberg, for example, we hold so-called town hall meetings to inform our employees about the steps we are taking towards implementing digitalization," said Beitinger.

Further Information

VDMA Fluid Power  |   VDMAimpulse 02-2017: "Machine Learning - A journey to the year 2030"   |   VDMAimpulse 02-2016: "Predictive maintenance for smart servicing"

Peter-Michael Synek, VDMA Fluid Power Association.