Until now a production plant had to fail before it was repaired or defective parts were replaced. Today a fault can be predicted before it even occurs. Predictive maintenance is the new catchphrase that saves an enormous amount of money and can open up new business models.
Predictive maintenance is the retrieval of maintenance information on the basis of real machine and production data with the aim of carrying out proactive maintenance of machines and plants before downtimes occur. Those who manage to predict and accordingly prevent impending machine failures can save themselves considerable stoppage costs. How can predictive maintenance be implemented?
1. Proactive maintenance through additional systems on the machine
There are several possibilities. On the one hand, predictive maintenance can take place by means of integrated machine monitoring. To do this, machines and industrial plants are equipped with a sufficient number of sensors that measure, for example, vibration, temperature or humidity and allow remote monitoring later on during operation. Defective components that would probably soon lead to plant downtime are thus identified irrespective of the usual maintenance times and can be replaced before damage actually occurs.
2. Condition monitoring with the help of the digital twin
Another way is to map the real machine behaviour with the help of a digital twin. The only difference is that the twin – i.e. the virtual counterpart of the plant – runs on a normal PC. In the ideal case the model is connected to the real machine controller and runs a defined number of cycles in advance. Failures on the virtual machine will also very probably occur on the real machine following the expiry of the defined time. The look-ahead period should be selected in accordance with the complexity of the plant and the magnitude of the stoppage costs in order to keep the downtimes as short as possible.
But how does one get the digital twin? machineering, for example, has created a platform – the simulation software iPhysics – with which a simulation model can be created simply and quickly from CAD data. Once connected to real controllers the virtual plant is brought to life and behaves like the real plant. In combination with manufacturer’s data on material wear, etc. a failure due to material fatigue can be predicted very reliably.
Predictive maintenance saves a lot of money
The main advantage of predictive maintenance can be reduced to a single catchphrase: cost-effectiveness. Regardless of the plants concerned: the savings potential is huge. The American railway company Union Pacific claims to save around 100 million dollars annually thanks to predictive maintenance. The maintenance of machines and plants is on its way into a new era. Even if this technology has only been used in a few companies so far, a clear trend in this direction is discernible.
Would you like to learn more? Just contact us, we’ll be glad to advise you!