How to Automate Image Processing in Pharma–Plants
An initial feasibility study showed that the optimum solution was the following sequence of steps:
- Intake of the granulate through a funnel system and transfer to a rotating glass disk with prior grain isolation in a vibrating channel;
- Creation of segments of the rotating disk, each to be monitored by optical cameras;
- Positioning of the cameras including the lighting in order to adequately capture the segments during disk rotation;
- Selection of suitable image processing hard- and software in order to identify out-of-spec grains per disk segment moving through; out-of-spec deviations relate to foreign matter, discoloration and grain size;
- Installation of a suction head, which aspires and removes the material from the respective segment in case of an out-of-spec signal from the image processing system;
- Definition of the operational optimum of the total installation depending on granulate throughput, disk rotation speed and response time of the defect identification and elimination system;
- Compilation of a comprehensive reference sample set representing the limits acceptable/not acceptable of the individual grains and serving as basis for qualification and operation of the installation;
- Development of an optimised and validated operational system in response to the prevailing requirements for quality and reliability combined with economic performance of the installation.
NNE Pharmaplan took over the responsibility for the whole engineering task (CD through DD) and handled the realisation activities as general contractor including the entire qualification process (from DQ to OQ). Components incl. system hardware and in particular the image processing part (cameras, lighting and the respective software) were procured in Switzerland and Southern Germany. Total realisation time from first feasibility studies to end PQ and hand-over to the operator was two years.
This article is protected by copyright. You want to use it for your own purpose? Infos can be found under www.mycontentfactory.de (ID: 42438236 / Pumps & Compressors)