Fault Diagnostics in Chemical Semibatch Reactors Using Neural Networks


Fault Diagnostics in Chemical Semibatch Reactors Using Neural Networks

Hessel, G.; Schmitt, W.; van der Vorst, K.; Weiß, F.-P.; Neumann, J.; Schlüter, S.

This paper presents a neural-network approach to early identifying dangerous states in chemical semibatch reactors. Data sets which were supplied both from a process simulator and from measurements in a laboratory reactor were used to train and test neural networks and a fuzzy pattern classifier for different normal and faulty states. Three-layer perceptron networks were found to be best suited for classifying different normal and abnormal process states. Even multiple fault states can be recognized by the perceptron network correctly.

  • Lecture (Conference)
    Proc. of the 5th European Congress on Intelligent Techniques and Soft Computing EUFIT 97, Aachen, Germany, September 8 - 11, 1997, pp. 1704 - 1708
  • Contribution to proceedings
    Proc. of the 5th European Congress on Intelligent Techniques and Soft Computing EUFIT 97, Aachen, Germany, September 8 - 11, 1997, pp. 1704 - 1708

Permalink: https://www.hzdr.de/publications/Publ-1618
Publ.-Id: 1618