FermBase is a datamining tool for the analysis of batch processes.

FermBase is a tool to analyse data from completed batch processes. Data from process computers from different plants are collected on a regular basis and processed for use with the FermBase application. Within FermBase the batch data can be analysed. The following methods are available:

  • A scatter plot with a moving average to compare performance indicator of batches over time
  • A trend analysis tool
  • Histogram and cumulative probability plot
  • A clustered data report in which parameters are clustered for a given period
  • Subset comparison in which batches can be grouped according to user configurable values
  • Ranked correlation, Partial Least Squares, Stepwise regression analysis
  • Anova
  • A searchable and filterable database

All graphs are interactive: data points are clickable to provide extra information, the graphs are zoomable and can be exported for reporting.
The batch data and timeseries can be cleaned for outliers and can be exported to Excel.
An important use of FermBase is to determine the values of critical parameters to monitor during processing. For this purpose so-called NOBS (Normal Operation Batches) can be generated and exported to use for the realtime monitoring in plants.