Predictive Maintenance

Avoiding unplanned downtimes or failures in your own production or at customers' by means of Predictive Maintenance is a central idea of Industry 4.0. This means that machines or products report to you "on their own" and inform you that a failure is very likely to occur soon. This must happen in time so that the maintenance intervention that prevents the failure can still be planned or commissioned without any problems. The more precise the prediction, the shorter the maintenance can be carried out before the failure and the closer you are to the technically optimal maintenance time.

Predictive Maintenance is a promising concept and, thanks to digital transformation and rapid developments in the field of connectivity, it is also within the reach of medium-sized companies. With Predictive Maintenance, dead times can be minimized and the availability of plants and products can be increased in the most efficient way possible. But it can also optimize areas such as spare parts management or service technician planning.

Predictive Maintenance is usually based on a condition monitoring scenario in terms of the technological setup. This means that a system that collects data from machines and/or products and puts it into context already exists.

For a Predictive Maintenance scenario, this data must now be analysed, correlations recognised and finally transferred into machine learning algorithms. In order to develop clean algorithms, very large amounts of data are necessary. However, these are usually not sufficiently available in an initial situation. Therefore, it is normally imperative to collect and analyse relevant data over a certain period of time as part of a condition monitoring system. A certain way out can be the qualification of data records by employees. If they use their experience to interpret measured values, it may be possible to arrive at reliable forecasts more quickly.

In principle, Predictive Maintenance should distinguish between concepts for in-house production (internal scenario) and for products and systems used by customers (external scenario), as different processes are affected in each case and different connectivity solutions are required.

Implementation of internal scenario: You can reach your goal comparatively quickly - because it is based on a ready-made solution - if you implement your predictive scenario with the SAP Industry Cloud Solution All for One Smart Factory as the foundation. This can provide you with the necessary data setup or a condition monitoring scenario as a by-product. Of course, we also support you with individual projects.

Implementation of external scenario: Here, too, we offer you a relatively standardised approach based on the setup underlying our use cases for mechanical engineering. Smart Products can provide an alternative basis.

Although Predictive Maintenance is sometimes a more complex undertaking in the final stage, it should be noted that the "preliminary stages" - such as condition monitoring - also offer practical and, above all, quickly realizable added value. For example, with the help of data analytics tools, patterns and correlations can be uncovered, which can quickly be converted into process or product improvements. Predictive Maintenance is therefore not a long journey in which added value can only be achieved at the very end.