LeanPredict covers all important cases of machine failure. It consists of a modular end-to-end solution equipped with various sensor types such as lasers, high-speed cameras, acoustic and vibration sensors, temperature sensors and others.

The data from the sensors and the various log files pass through an AI process (Artificial Intelligence)where anomalies of any kind are detected. This provides users with prognostic key figures via dashboards and alarm systems.

Also, the standard interfaces (connectors) can be used to connect various systems such as MES, SCADA, SPS (controls), ERPs and various conventional maintenance software.

LeanPredict offers a worldwide unique framework product for predictive maintenance.


Our engineers have a high level of expertise in a wide range of sensor technology for predictive maintenance. Depending on the case of damage they are used:

Image sensors (vision)

Laser 3D triangulation can be used to measure bends, displacements and cracks in the sub-millimetre range. This method is a very accurate method to detect geometric misalignments such as bends and cracks at an early stage.

It is also possible to detect rapid changes with cameras. For this we use specific sensors of our partner Insightness. By a patented compression technology of the data directly at the sensor changes in the millisecond range can be detected. So fast that accidents can be avoided.

Acoustic measurement technology

In principle, acoustics is a powerful, contact-free and at the same time very inexpensive measuring technique for detecting anomalies. Acoustics therefore has advantages over vibration measurement technology. However, the evaluation is complex, especially if there are ambient noises. In our projects we have built up sound know-how on this measurement technology.

Vibration measurement technology

The vibration measuring technique is the right measuring technique to detect different wear conditions. This works on bearings, gears, but also on chains and belts.

Without sensor technology

Sensors cost. Therefore, the primary goal is to carry out predictive maintenance without additional sensor technology. For example, in many cases a specific anomaly can be inferred from the power consumption of the motors in combination with rotary encoders.

With less sensor technology

There is a potential with machine learning to replace expensive sensor technology with a combination of inexpensive sensor technology. This is referred to as virtual sensors. Here,too, we were able to build up in-depth know-how in the projects.