Predictive maintenance is the logical continuation of condition monitoring, which has long been integrated into many modern industrial systems as a further development of classic operating hours recording. While condition monitoring only enables the detection of wear and tear, predictive maintenance ideally allows maintenance appointments to be scheduled well in advance. This results in higher system availability and reduced costs.

Especially with smaller geared motors, which are installed in large numbers in intralogistics installations, the use of additional 'real' sensors for condition monitoring - as used in industrial gearboxes - is often too cost-intensive. With this in mind, Nord Drivesystems has developed a concept that can use the 'vital data' of the drive or 'virtual' sensors to calculate the status of the system. Such a concept offers an economical implementation of predictive maintenance, especially for systems with many small drives.

Using virtual sensors based on mathematical algorithms and the PLC integrated in the inverters, the frequency inverters can pre-process the internal status values to calculate variables that can only be measured with great effort. For example, the electrical data is used to calculate the drive power, which, in combination with other parameters and gear oil data, provides a sufficiently accurate indication of the oil service life. This makes the degree of utilization of the transmission oil and thus the expected oil change date accessible. Depending on the load, utilization and installation of the respective geared motor, these dates can differ significantly even within a system with geared motors of the same age.

Another scenario plans something similar for the prediction of the wear condition and the ideal maintenance date by means of a target and actual comparison using an algorithm: In a learning phase, the electrical data is determined on the new conveyor system in an unloaded and loaded state and defined as reference values. If these are exceeded later in real operation, the frequency inverter recognizes that something has changed in the mechanical system. This may be due to increased friction, wear, a damaged bearing or gearbox, or foreign objects (packaging material, adhesive tape) that have become trapped. If the mathematical correlations of the system are known and transferred to validated intelligent algorithms for data evaluation, predictive maintenance can be carried out for the drive technology even without real sensors.

Healthcare for large gearboxes

Industrial gearboxes are the heavyweights in drive technology and have to transmit high torques. Small, unnoticed defects can quickly lead to damage due to the large forces acting on them. This would not only be expensive, but also fatal: important system components would come to a standstill until a replacement is delivered and installed.

It is therefore also advisable to rely on condition monitoring and predictive maintenance for industrial gearboxes in order to achieve maximum system availability with high efficiency. Intelligent frequency inverters with an integrated PLC that can make autonomous decisions are also a prerequisite in this case. However, the costs of physical sensors are not as high in relation to the gearbox costs, which is why real temperature and vibration sensors also make sense.

Pilot conveyor system at Nord Drivesystems in Bargteheide: With the cloud connection, a technician can evaluate the status data of the drives in a clearly structured manner, regardless of location.

© Nord Drivesystems

Vibration sensors in particular offer a number of advantages in this context. For example, there are detailed manufacturer databases for the bearings installed in industrial gearboxes, which contain the characteristic vibration frequencies of the inner ring, outer ring and rolling elements of each bearing type. The gear mesh frequency, the bearing frequency and the respective speed are also known. The individual frequencies can therefore be clearly identified and assigned.

Using the time signals or an FFT analysis, the frequency spectrum can be examined on this basis in order to clearly identify the causes of the vibrations that occur. The FFT analysis (Fast Fourier Transformation) is an algorithm for the efficient calculation of the discrete Fourier transformation (DFT). It can be used to break down and analyze a digital signal into its frequency components. The determination of the frequency components, amplitudes and phases of the oscillations and their comparison with the oscillation databases thus allow a detailed status diagnosis. In this way, not only can the ideal or necessary maintenance time be calculated, but it also becomes clear where the fault may lie and which spare parts are required.

Cloud connection for networked service

In principle, a cloud connection is easy to implement, even when retrofitting existing systems. All drives have their own IP address via which they can be reached using a router. The data obtained through condition monitoring and predictive maintenance for each individual drive in the system can be queried without interfering with the device control or the software.

Useful for large gearboxes: Check the gearing using endoscopy.

© Nord Drivesystems

The intelligent drive components transmit the values to a secure cloud via an internet gateway, for example. There they are available for evaluation with filter and analysis tools.

This means that a technician can analyze the data at any other location in a clearly structured browser-based web interface. It is possible to determine what information the drive expert needs and what information is important for the service technician.

Nord Drivesystems is currently developing the relevant know-how for all these new technologies, in some cases in cooperation with partners, in order to implement them with users in their systems in the future.

Author: Jörg Niermann is Head of Marketing at Nord Drivesystems.