Automotive suppliers are under enormous pressure. Unscheduled production downtimes cause high costs. This is all the more true when production processes are as closely interlinked as they are at Buderus Guss. The European market leader for car brake discs produces 20 million cast parts a year in Breidenbach, Hesse, which have to arrive just-in-time at the international manufacturers.

High system availability is a must

The first link in the Buderus Guss production chain is formed by the five highly automated molding lines. If their hydraulic drive fails, costs of several thousand euros are incurred every hour. It becomes significantly more expensive if the buffer of blanks for further mechanical processing of the brake discs runs empty. To prevent this, the forming systems must achieve the highest possible availability. Until now, the company has only been able to achieve this with regular maintenance work. "The pump, which weighs around 70 kg and is installed at a height of 3.50 m, was regularly replaced as a preventative measure," explains Thomas Heck, Head of Technical Maintenance at Buderus Guss. "In the event of an unplanned breakdown, the emergency repair, including scaffolding and special lifting tools, would have resulted in production downtime of up to eight hours."

The ODiN (Online Diagnostics Network) service package from Bosch Rexroth uses the interaction of sensor technology, cloud-based applications and machine learning methods to detect critical faults or significant changes to the normal operating status in advance and warn of a probable failure in good time. (Image source: Bosch Rexroth AG)

In order to better assess the risk of pump failure and thus reduce it in the future, Buderus Guss initially implemented various measures for local condition monitoring with the help of Bosch Rexroth. For example, sensors were installed in the filter system to continuously monitor the fluid flow. However, the preventive pump replacement every six months and the high spare parts costs remained, even though the service life of the pump could be significantly longer depending on the utilization. "This problem could only be solved by a predictive warning system, which made it possible to schedule a planned replacement during the non-production maintenance window at the weekend if action was required," reports Thomas Heck. "However, the necessary warning time until the next planned shutdown was not feasible with the previous methods."

The decisive step towards a condition-based maintenance strategy that maximizes the service life of the motor and pump while minimizing the risk of failure was only achieved with the help of the Online Diagnostics Network (Odin). The service package uses the interaction of sensor technology, cloud-based applications and machine learning methods to detect critical faults or significant changes to the normal operating status in advance and warn of a probable failure in good time. At Buderus Guss, the sensor system installed by Bosch Rexroth now also records the vibration of the pump and the oil quality. The latter also serves as an indicator for the condition of the valves and for precisely scheduling the filter change.