Predictive maintenance at Bosch minimises downtime

Predictive maintenance at Bosch minimises downtime
Machines idle for maintenance cause downtime and high costs. However, maintenance can take different forms, from corrective to proactive and even predictive. Corrective maintenance is the least efficient: here, action is taken only when defects or outages occur, leading to disrupted production schedules and even possible damage to production systems. Many companies therefore foresee most maintenance proactively, scheduling maintenance at fixed times or after a certain period of use, trying to predict the lifetime of components. However, this form of maintenance is also not flawless and in many cases happens too early or too late, rarely just in time.
The industry therefore needs more efficient methods. Via predictive maintenance, we want to try to predict defects even better at Flanders Make. This involves collecting extensive information about a machine (component), for instance by measuring vibrations, noise levels, temperatures or analysing video images. By continuously monitoring the machine, we can predict more and more accurately when certain parts need to be replaced. This way, a company only needs to stop production for maintenance at the inevitable moments.
"We predict material failure up to two weeks in advance by monitoring and analysing gear vibrations."
Bosch
What did Flanders Make do for Bosch?
Bosch in Tienen produces wiper blades and wiper arms. For its wiper blade production line, the company wanted to optimise uptime by switching to predictive maintenance. Toothed belts are the critical component here.
We conducted a feasibility study to select sensors and develop the necessary algorithms to monitor the vibrations - and consequently the condition - of the timing belts. The final approach using accelerometers was validated on a number of production lines, with consistently good results. For example, Bosch was able to detect the failure of timing belts two weeks in advance.