Golden Batch Analysis: what separates good penicillin batches from bad ones
Historical data from 100 fermentation batches, split into high- and low-performing groups. Most operating variables looked the same; PAA and substrate concentration did not.
Blog
Insights
Notes on soft sensors, batch analysis and process control, written by the engineers who build them.
-
Case study
-
Case study
Soft sensor for pharmaceutical processes
Penicillin and PAA are measured in the lab every 12 hours. Hybrid models with physics-derived variables close that gap, raising R² from 0.95 to 0.98.
-
Guide
Soft sensors: what they are and how to build one
A soft sensor estimates an unmeasured process variable from existing plant measurements. Where they are used, the five-step development workflow, and where the RPDS tool helps.
-
Process control
Process control and optimization
From the Watt governor to model predictive control: why precise control matters in the chemical industry and what RPDS does in this area.
-
Maintenance
Predictive maintenance in the chemical industry
How maintenance moved from reactive and scheduled to data-driven, and the workflow for detecting anomalies, classifying faults and estimating remaining useful life.