Datasets

Publicly available datasets for testing (adaptive) soft sensors

No Process Data Type No. examples Dimensions Target Reference
1 Polymerisation real batch 4500-7500 per batch 34 Activity/viscosity [2]
2 Fermentation simulated batch variable 6 9 options [3]
3 Polymerisation simulated continuous 8687 15 catalyst activation [4]
4 real continuous
1440 per fault, 21 faults
52 n.a. [5]
5 real continuous 2393 7 Butane concentration [6]
6 Butane concentration real continuous 10080 5 concentration [6]
7 Distilation column real continuous 500-900 per fault, 10 faults 19 n.a. [7]

Table source: adapted from [1]

[1] Kadlec P., Grbic R., Gabrys B. Review of adaptation mechanisms for data-driven soft sensors. Computers and Chemical Engineering. 35(1):1-24, 2011.
[2] Facco, P., Bezzo, F., & Barolo, M. Nearest-neighbor method for the automatic maintenance of multivariate statistical soft sensors in batch processing. Industrial and Engineering Chemistry Research, 49(5), 2336-2347, 2010.
[3] Birol, G., Undey, C., & Cinar, A. A modular simulation package for fedbatch fermentation: Penicillin production. Computers and Chemical Engineering, 26(11), 1553–1565, 2002.
[4] NISIS'06 competition, also Kadlec, P., Gabrys, B. Local learning-based adaptive soft sensor for catalyst activation prediction. AIChE Journal. 57(5), pp. 1288-1301, 2011.
[5] Chiang, L. H., Russell, E., & Braatz, R. D.. Fault detection and diagnosis in industrial systems. Springer, 2001.
[6] Fortuna, L. Soft sensors for monitoring and control of industrial processes. London: Springer Verlag, 2001.
[7] Seng Ng, Y., & Srinivasan, R. Multi-agent based collaborative fault detection and identification in chemical processes. Engineering Applications of Artificial Intelligence, 23(6), 934-949, 2010.

Please contact the INFER project coordinator Prof. Bogdan Gabrys if you need any further information on the data sets.