GermanyChemicals and materials2025 – ongoingFUCHS
Alchemite machine learning for lubricant formulation
FUCHS tested 7 machine-learning candidates and found 2 that met its lubricant performance target
Problem
FUCHS needed to adapt an existing lubricant to aluminium forging. Complex formulation projects required repeated experiments across unfamiliar materials.
Approach
Researchers trained Alchemite on 24 historic formulations. The model proposed 30 candidates across successive experimental cycles. A chemist selected 7 for testing, and 2 met the target of a 50% performance improvement.
Results
Takeaway
The model narrowed the experiments while chemists retained control over testing. The specific performance measure and elapsed time savings were not disclosed.
Categories
- Industry
- Chemicals and materials
- Business function
- IT and R&D
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and MLOptimization
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Intellegens described FUCHS using Alchemite to guide lubricant experiments. It reported that 2 tested candidates met the target and that 1 was being pursued further.
Full details are in the original sources.
