Predictive Maintenance in practice

We compare MEMS and piezo-based sensors and show, that for example MEMS are less suited for low RPM applications, i.e. below 500 RPM. It also shows the risk of using RMS data only, like most low-cost IoT systems do, to determine the health state of a bearing.

Predictive maintenance (PM) lends itself to bearings and gears because defects appear deterministic. In practice, it has been shown that the problems do not start in the analysis, where artificial intelligence is used today, but in the extraction of the raw data. Lost information in the data cannot be recovered later.

Defects in gearboxes or rolling bearings would like to be detected at an early stage so that repairs or possible replacements can be planned and carried out. If the monitoring system can only measure a defect when it can be felt or heard, then the risk of failure increases massively. In order to be able to use predictive or condition-based maintenance, e.g. to optimize lubrication intervals, we consider it essential to use high-resolution raw data. Furthermore, this can reduce the false alarm rate and increase defect detection. Of course, this requires the use of good sensor technology and the appropriate electronics for data digitization. This does not have to be expensive, as we have proven, but must meet the expectations of PM.

You can find more information (in German) here.


More interesting posts

Saved in your Bookmarks

We use cookies on our website. Some of them are essential, while others help us to improve this website. We are grateful if you accept them. More info