Munich, 11 March 2021
The digitalisation of industrial production continues to advance. One current development is predictive maintenance, which means predictive maintenance and needs-based maintenance. This is based on the processing of process and machine data in real time and thus makes it possible to anticipate the unexpected failure of plants. At "acatech on Tuesday" on 9 March, the speakers gave an overview of the benefits of predictive maintenance and the status quo in German companies. The experts agreed that predictive maintenance is still far too often a thing of the future. The event was held in cooperation with the VDI Bezirksverein München, Ober- und Niederbayern e.V.
At the beginning of the event, acatech President Dieter Spath briefly introduced the panel and guests to the topic: Predictive maintenance - that means looking at something as a precaution and at the same time checking how production is running. When is the next maintenance due, when would it have the least impact on the production process? This is all based on processing process and machine data in real time in order to anticipate an unexpected failure and thus better plan spare parts and personnel for maintenance work and thus save costs, said Dieter Spath. Andreas Wüllner, Chairman of the VDI Bezirksverein München, Ober- und Niederbayern e.V., then introduced the Verein Deutscher Ingenieure" (VDI), the largest technical and scientific association in Germany, and then passed the floor to his fellow board member Christa Holzenkamp, who moderated the event and guided the audience through the evening.
The status quo of predictive maintenance was then described by acatech member Michael Henke, Director of the Fraunhofer Institute for Material Flow and Logistics in Dortmund. Michael Henke began by pointing out current problem areas that make predictive maintenance difficult: Production and maintenance plan separately, production is relatively inflexible, know-how is strongly tied to individuals, the spare parts system is not structured and there is a predominantly reactive maintenance, which is primarily seen as a cost generator. In the future, predictive maintenance should lead to joint planning, availability-oriented maintenance, more flexible action and reaction to changes, better knowledge and spare parts management. This should make the contribution of maintenance to the company's success visible. Michael Henke named the lack of capacities in the IT sector and thus the lack of know-how in the companies as hurdles that delay or impede the use of predictive maintenance. A cost-benefit ratio that has not yet been clarified means that predictive maintenance has not yet been implemented across the board in corporate practice. In research, it is an interdisciplinary field of research on which departments such as business administration, mechanical engineering and data science are working in equal measure. Michael Henke concludes: Without maintenance, there will be no smart factory. That is why we will increasingly need a smart and predictive maintenance understanding and approaches in practice in the future.
Markus Loinig, Senzoro GmbH, Vienna, deals with the question of customers and maintenance staff about the right time to replace a component. His start-up develops solutions for predictive maintenance and condition monitoring - based on the question of which technology optimally records the condition of a system and makes reliable predictions for the future. According to studies, ultrasound is the fastest way to detect bearing damages. Moreover, compared to other technologies such as oil, vibration, infrared and machine parameters, it is characterised by little background noise and thus provides very "pure" data. Using an example from the wood industry, Markus Loinig showed how Senzoro's system can measure the condition of roller bearings using mobile broadband ultrasonic sensors. AI then processes the data and outputs a health score. This reduces maintenance costs by predicting the remaining service life of roller bearings and replacing them according to their condition.
Andreas Wüllner began the subsequent discussion with a sobering insight into industrial practice. Although unplanned downtimes were horror scenarios, maintenance would be seen as a necessary evil, because the focus is on the cost side and productivity. All too often, maintenance staff are not integrated enough into the organisation and work in the company alongside and not with production. Despite its great importance, the topic of maintenance has not yet reached management in two-thirds of the companies. There is a lot of catching up to do here.
Dieter Spath reported on an industrial sector that serves as a role model. The construction machinery industry, he said, has achieved through the connectivity of its products that mechanics can keep an eye on remotely monitored machines.
Michael Henke was able to confirm that there is still too much surprise in maintenance. Predictive maintenance should not be a dream of the future, even if it often seems so in reality, because although the benefit compared to reactive maintenance is obvious, the topic is not yet widely anchored. He called for the various entities in the company to be brought together and for data to no longer disappear into silos. He is convinced that maintenance is one of the main drivers of digital transformation.
Who commissions it depends on the size of the company and is management-dependent, Markus Loinig said. The decision as to whether the introduction of predictive maintenance is worthwhile is highly dependent on the company and also depends on whether failure components can be modelled cost-efficiently. If there are no follow-up costs for failures, reactive maintenance is preferred.
In order to bring the employees along on the path to predictive maintenance, Michael Henke concluded the committed discussion with an appeal for the development of further training offers, IT capacities and improved network communication. You have to start at school, teach children about STEM subjects and show them how cool the topic of maintenance is.
Translated into English by Senzoro GmbH
