ALL libraries (COBIB.SI union bibliographic/catalogue database)
  • Hybrid modeling approach in chemistry
    Potočnik, Primož, 1969- ...
    A hybrid approach to modeling of chemical processes is proposed. The hybrid model consists of an analytical model and a neural network, connected in parallel. The purpose of a neural network is to ... improve the prediction accuracy in the domain where enough measured data exist. A radial basis function type neural network is applied for this task due to its local approximation property. A case study "Modeling of the liquid-phase methanol synthesis" is presented where an analytical and a hybrid model are constructed for the prediction of methanol production rate. Significant improvement in the prediction accuracy is demonstrated by using the hybrid modeling approach compared to the analytical one.
    Type of material - conference contribution
    Publish date - 2001
    Language - english
    COBISS.SI-ID - 4591131