At a glance
The principle of “Predictive maintenance” uses previously collected data on the maintenance and repair of machines to make predictions about the condition of these machines. The forecast also works while the machines are in operation and can therefore be used continuously.
Prerequisites for predictive maintenance
At the core of predictive maintenance are accurate, forward-looking statements on how the condition of plants and individual machines will develop. For this purpose, the system uses a series of sensors that collect data in real time on the condition of the equipment, but also on its surroundings. This data is collected in a central database and then analyzed. From this analyzed data, the system can then calculate how likely it is that which event will occur when, e.g. when a wear part is worn out. Networking between machines, sensors and the predictive maintenance system is done, for example, via the Internet of Things .
Predictive Maintenance with Big Data
The more data available for the forecasts, the better they will be. This not only means that data must be collected at as many measuring points as possible in and around the plants: The analysis and long-term storage of these large amounts of data must also be planned. This is because forecasts only become truly effective when historical data can be used for analysis. This is the only way, for example, to identify certain patterns, make reliable statements or distinguish unusual situations from everyday life. Big data technologies provide the necessary capacities to operate the system with high performance and stability.
Comparison with traditional maintenance methods
Traditional maintenance methods often rely on a fixed schedule or on a routine adapted to the workflow, according to which certain tasks have to be performed for the maintenance and servicing of machines. The aim of this method is to maintain the machines preventively - i.e. as far as possible before wear and tear and breakdowns can occur, which could disrupt operations. Predictive maintenance, on the other hand, tries to determine as precisely as possible when maintenance or inspection will be necessary according to the actual utilization and wear of the machine.
Successful predictive maintenance can help determine the best time to service machines. This has the following advantages:
The constant analysis of conditions and the prognosis of possible events help to minimize machine failures: Instead of relying as usual on fixed maintenance intervals or even servicing the equipment in the event of damage, potential problems are detected and reported at an early stage. It is therefore less common for a machine to come to a standstill unexpectedly due to a defect. In addition, it can be ensured in good time that the spare parts that are likely to be required are in stock.
Human labor is one of the most expensive resources in industry. By having sufficient data and less or no fixed preventive maintenance intervals, the freed up working time can be invested elsewhere; the maintenance and repair of the machine becomes cheaper because only work takes place when the condition of the machine actually requires it.