Success Story Predictive Maintenance

Pharmaceuticals

In order to manufacture a drug, a number of process steps must be performed precisely in sequence to obtain a final product that can be accepted. Even slight deviations in processing temperatures and times can render an entire batch unusable. Using artificial intelligence, differences in the compliance of processing parameters can be detected early and reliably.

The Case

• During production, different raw materials must be brought into relation to each other under precisely defined conditions
• If a process step takes too long, or if temperature corridors are not strictly observed, an entire batch can be rendered unusable
• One plant can produce dozens of different end products. For each drug, though, different processing steps and different parameters apply
• Certain process steps, such as drying, are condition-dependent and can take a variable time
• Inconsistencies in processing can occur due to shift changes or malfunctions
• If, for example, steam has escaped through a small leak, the air humidity rises. Without timely correction, a large loss would be the result