Reducing WWTP models

The existing waste water treatment plant models, like the ASM models, are quite complex and take a long time to solve numerically.
Consequently, doing a sensitivity or risk analysis requires a lot of simulations and hence a lot of computation time.
The technique used to reduce the models was the building of artifical neural networks (ANN) which are claimed to have a good interpolation capacity when trained properly. Training was done with data generated by the full complex ASM2d model. However,the neural network led to unstable predictions as errors were accumulating. Another model reduction technique was used instead, namely a knowledge based model reduction, in which knowledge is used to eliminate or simplify processes.

Project duration: 1999-03-01 - 1999-05-31

Project home page

No website available


Jurgen Meirlaen
Department of Applied Mathematics, Biometrics and Process Control
Coupure Links 653
9000 Gent
Tel: +32(0)9 264 59 37
Fax: +32(0)9 264 62 20

Last update: 01 december 2008,

  zoek print log in nederlands