SBMBR 2018 Functional Fault Detection Study

A dataset from a demonstration-scale SB-MBR system in Mines Park at Colorado School of Mines from June 01, 2018 at 12:04 AM to September 5, 2018 at 1:40 PM. The method of ultrafiltration used in the system alternates between periods of flux permeation and short backwashes, such that the flux permeation periods can be modeled as distinct functions over time. Process faults can result in excessive trans-membrane pressure (TMP), so this data can be used to detect faults by observing the ultrafiltration cycles with functional data analysis techniques. One fault that may be detected is a drift fault beginning on August 26, 2018. This dataset contains a data frame with 10,518 rows and 61 columns. Each row represents one ultrafiltration cycle where the first column represents the starting date and time for the ultrafiltration cycle, and the remaining 60 columns represent measurements of TMP within the ultrafiltration cycle, equally spaced at 5 seconds apart.