Predictive maintenance, based on an intelligent monitoring system, predicts and prevents possible equipment failures. Data analytics of transportation equipment allow early detection of problems, thus helping to maintain optimal flow in the vehicle production process.
The system using artificial intelligence (AI) prevents an average of around 500 minutes of downtime per year in vehicle assembly at the Regensburg plant.
Data analysis for faster response to potential disruptions
For assembly at the BMW Group factory in Regensburg, vehicles are generally placed on mobile loading docks or chain-driven skid systems that pass through the production halls. Any technical error in the state-of-the-art conveyor systems can lead to a standstill in the assembly lines.
This increases both maintenance and cost requirements. To prevent this, the Regensburg plant’s innovation team developed a system that can detect potential technical problems early – and thus avoid any production losses.
Affected parts of the transport element can be removed from the assembly line and repaired away from production. The advantage is that the monitoring system does not require additional sensors or equipment, but evaluates the existing data from the installed components and the control system of the transport elements. If abnormalities are detected, an alarm sounds.
For example, the loading docks used to transport vehicles on the assembly line send various data to the transporter’s control system, which is then transmitted via it and the factory’s control unit to the BMW Group’s cloud-based preventive maintenance platform. From there the analysis begins:
The algorithm is constantly looking for irregularities, such as fluctuations in energy consumption, malfunctions in conveyor movements or barcodes that are not sufficiently legible, which could cause problems. If these are detected, the maintenance control center receives a warning message, which the maintenance technician on duty is informed about. Monitoring screens in the control center operate 24 hours a day, allowing shift controllers to respond quickly to any type of fault report and remove the affected vehicle from the process cycle.
Implementation – powered by artificial intelligence
The system was standardized in cooperation with the central management of the BMW Group production unit and other facilities, so that it can be immediately implemented in other factories of the group around the world. This approach is also cost-effective because no additional sensors are required, so the cost is only for storage and computing power.
Machine learning models developed in-house were applied to the system, which uses so-called heat maps with various color codes for the various anomalies to visualize the model’s findings. This allows different failure patterns in various components to be mapped and engineers to respond to them in a targeted manner.
Based on these practical findings, the algorithms are continuously improved and refined. The team is currently in the process of connecting additional installations, optimizing the system, and incorporating recommended actions into error messages. This could, for example, indicate similar problems that have occurred in a system making it easier for maintenance technicians to deal with them – for example, if an impeller on a transport trolley is faulty. Optimal predictive maintenance, in addition to the financial benefit, helps to deliver the planned quantity of vehicles on time – which greatly reduces stress on production.
Next goal: Predictability
Over the past six years, data-driven technology monitoring of transportation systems has been implemented. Today, around 80% of major assembly lines are already tracked in this way. At the BMW Group plant in Regensburg, a vehicle rolls off the assembly line approximately every minute – every 57 seconds to be exact – and the system is already being used in transport systems at the plant sites in Dingolfing, Leipzig and Berlin.
The goal is to further utilize the capabilities of artificial intelligence, with the system learning to estimate how much time is left from the moment of fault detection until a possible shutdown. This would help the technicians to decide how soon to perform the maintenance tasks and also to prioritize their actions if necessary. This period is also looking at the possibility of using the system in equipment used to fill vehicles with brake fluid and coolant, for example.
