Predictive Maintenance in Pharma
Identify the probability of machine breakdowns and increase your product quality
Challenges
Leading pharmaceutical company
Extremely high packing standards mean product spoilage is very costly
Machine breakdowns substantially reduce overall plant efficiency
Client wanted to predict probability of machine breakdowns and quality defects
Solutions
Combined data from production line sensors with the Bill of Materials (BOM), input-material quality measures and machine operator team
Built predictive Machine Learning models that successfully explained the probability of a machine breakdown and quality defect
Developed combinatorial optimization method, inspired by genetic algorithms, providing the machine operator with the optimal machine parameters for the given product
Built an app to visualise the optimal machine parameters directly on the operator’s tablet
Values
Team running the production and packaging lines gain a high degree of confidence in the machine parameters suggested by the app
Significantly reduced machine breakdowns and quality defects with the algorithm continuing to learn from new equipment, products and materials
Gained additional insights on the ideal composition of skills for the employees operating the production line
Roles
Machine Learning Engineer
Technologies
Describe your image |
---|
Describe your image |
Describe your image |
Describe your image |
Describe your image |
Describe your image |
Sectors
Health & Pharma, Manufacturing