Cognitive sensor and cognitive control concept in pharma use case

The development of the cognitive sensor solutions and the cognitive control concept in the pharmaceutical use case has progressed as planned: The sensor for measuring the active pharmaceutical ingredient (API) content in the granules has been implemented and tested on RCPE’s GEA ConsiGma CTL25 manufacturing line. After designing and 3D-printing of the required parts for the sensor interface (see the CAD drawing in Figure 1 and the 3D printed prototype in Figure 2), the respective algorithms for actuation of the sample holder, the computation of the concentration values from the captured spectral data and providing the result via OPC UA interfaces has been implemented in Python.

Figure 1: 3D-printed sensor interface

Figure 2: 3D-printed sensor interface

The development of the cognitive control concept for the granulation step has been pursued. Based on information provided by the cognitive granule quality sensor, settings have been automatically adjusted. This concept allows the precise production of granules of desired size. Compared to conventional operation, which uses constant granulation settings, the granule size can remain closer to its desired value, which in turn improves the robustness of the entire process depicted in Figure 3.

Figure 3: Control concept of the granulation step has been implemented on the manufacturing line

The next steps in the project cover testing of the developed solutions by means of experiments on the manufacturing line. Additionally, the created algorithms will be partly re-written in Python and transferred to the project partner Engineering, who will then integrate them into the cognitive automation platform (CAP).