Silk Purse or Sow’s Ear?
Process Data, Information and Knowledge
The aim of this course is for you to learn how to extract information, and perhaps even knowledge, from the plethora of process data generated by SCADA, DCS and plant historians.
Issues addressed
- Process variables worth measuring and why
- Ensuring measurements contain the information you want
- Averaging and filtering – when to use them
- Monitoring, detection and diagnosis
- Encapsulation of process knowledge
- Cause-and-effect (lies, damned lies and statistics)
- The pitfalls of data models
What do you get?
- Real plant data and exercises
- Course notes developed by leading researchers and practitioners
- Interactive discussion groups
Programme
Day 1 – Process Data
- Characteristics of process data
- Simple statistics refresher
- Variables worth measuring
- Measurements worth analyzing
- Simple signal processing refresher
Day 2 – Information
- Outliers, noise and filtering
- Detecting shifts and trends (SQC and more)
- Detecting oscillations
- Correlation and principal components (PCA)
Day 3 – Knowledge
- Correlation and regression
- Using and abusing cause-and-effect
- Simple time-series modeling
- Neural networks and pattern recognition