Predictive maintenance techniques are intended to help determine the condition of the devices that are in operation. This is used to estimate when maintenance should be carried out. This approach promises cost savings compared to routine or long-term preventive maintenance, as tasks are only performed when they are actually necessary. They are thus viewed as condition-based maintenance that is carried out when an object or a system is in a poor condition.
The main promise of predictive maintenance is to allow convenient planning of corrective maintenance and avoid unexpected equipment failures. The key to this is the “correct” service life of the devices, increased system safety, fewer accidents and optimized spare parts handling. Predictive maintenance differs from preventive maintenance in that it relies on the actual condition of the equipment rather than its average life.
Some of the main components required to implement predictive maintenance are data collection and preprocessing. Furthermore, the early detection of errors, the prediction of the time to failure, maintenance planning and resource optimization. Predictive maintenance was also seen as one of the driving forces behind improving productivity and one of the ways to achieve “just-in-time” manufacturing.
The prediction of conditions in systems can be realized using large amounts of data, which are usually recorded non-destructively with sensors. Examples of such data are camera images, temperature and humidity values, acoustic signals or vibration data from components or systems. With intelligent machine learning algorithms, this data can be analyzed and conclusions can be drawn about the current or future state. In order to enable continuous monitoring, this sensor data must be recorded at regular intervals and analyzed using the evaluation algorithms.
Predictive maintenance is already used in many areas. For example, speeds, noises, temperatures and vibrations in systems and engines are monitored in order to detect irregularities at an early stage.
In this way, blockages in nozzles or other components in printing systems can be detected at an early stage. With suitable measures failures can be prevented or major maintenance interventions can be planned.