BMW Group Plants

Manufacture: Data Entry Automation for one of the BMW Group Plants

 

Outcome:

•  4600 working hours per year
•  31% increase in output capacity
•  23% cost reduction.

 

How?

 

Situation:

In a plant where the vehicles are assembled throughout 100 production points, where at each point parts are fitted by line workers. Each production point fits a different part, such as door panels, tachometer, etc. Each production point has its own cost centre, which is charged in case damage occurs during their process.

Challenge:

Defects are not always noticed or reported by line workers and often, are only noticed during quality control procedures or at a later stage down the production line, however, when this is the case, workers don’t know from which production point the defects originate from and are therefore assigned to the wrong cost centre. If a cost centre gets a wrongly assigned defect, then they are responsible to identify where the defect originates. Collectively production points wasted 90+ hours weekly on correcting the cost centre selections in their legacy systems.

Solution:

A predictive model was developed and integrated into their system. The predictive can predict where the damage originates with a 100% accuracy based on defect description. Data with correct entries and workstation descriptions was used to train and refine the model.