Problem
A pharmaceutical cleanroom restricts entry by design. Checking equipment or discussing an issue means gowning up and clearing an entry procedure, and bringing in an outside expert or staff from another site is harder still. Production data, cleanroom environmental readings and electronic documents lived in separate systems, so no single screen showed the situation. Equipment was staffed one operator per machine, tablet press by tablet press, coater by coater.
Data collected
Production records, process information, cleanroom environmental status and electronic documents — previously scattered across existing systems — were linked into the virtual factory. Equipment operating data was used for both remote control and quality prediction.
Modules applied
A virtual factory identical to the physical facility was built and combined with big data, AI and equipment control technology. An AI model trained on production information predicts quality and derives the process conditions for optimal yield and quality. Remote control was implemented for pharmaceutical production equipment including KILAN tablet presses and GLATT coaters. Developed jointly with Tobworks.
Result
Internal and external staff now collaborate in the virtual space through voice, text and integration with the messengers they already use. Production information, cleanroom status and electronic documents appear on one screen. Remote equipment control makes it possible for one person to operate multiple machines. In production use on site since March 2024.
A factory you cannot walk into
A pharmaceutical cleanroom is not open to everyone. That is the rule, and it should be. The problem is that the restriction becomes a restriction on collaboration.
When equipment shows warning signs and needs discussion, it is hard for the relevant people to look at the same thing while they talk. One person goes in, looks, comes out and relays what they saw — slow, and information is lost along the way. For an outside expert or staff from another site, harder again.
So we built a second factory — virtually
A metaverse factory goes a step beyond a smart factory. You build a virtual plant identical to the real one and connect live data to it. The physical site and the virtual space stay linked and act on each other.
Inside the virtual factory, equipment and processes can be viewed without an entry procedure. Internal and external staff gather in the same space, talk over voice and text, and keep using the messenger they already have. Lowering the threshold for discussion was this system’s first purpose.
Scattered information on one screen
Production information sat in the production system, cleanroom status in the environmental system, related documents somewhere else again. Understanding a situation meant moving between three places.
Linking that data into the virtual factory put production records, cleanroom status and electronic documents on a single screen.
Prediction and remote control
Two more layers went on top.
Quality prediction — an AI model trained on product production information predicts quality and derives the process conditions needed for optimal yield and quality.
Remote equipment control — KILAN tablet presses, GLATT coaters and other pharmaceutical production equipment are controlled remotely from the virtual factory. A structure that needed one person per machine becomes one where a single person runs several.
Where it stands
Developed for a pharmaceutical customer under the 2023 Metaverse Factory support program, and in production use on site since March 2024. Not a demonstration system — one that is actually being used.
References
- Chong Kun Dang joins the Metaverse Factory support programetoday · 2023-06-14