Qualitative simulation of chemical process systems: Steady-state analysis

Authors

  • Olayiwola O. Oyeleye,

    1. Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139
    Search for more papers by this author
  • Mark A. Kramer

    Corresponding author
    1. Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139
    • Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139
    Search for more papers by this author

Abstract

A method of qualitative simulation for continuous process systems that predicts the steady-state measurement patterns generated as a result of process malfunctions is presented. The method is based on qualitative versions of the steady-state process equations and requires no numerical information beyond the signs and relative values of certain groups of parameters. Previous methods of qualitative simulation (Umeda et al., 1980; de Kleer and Brown, 1984; Forbus, 1984; Kuipers, 1986) have tended to generate multiple solutions that do not coincide with any observable system behaviors. In the current method, spurious solutions are reduced using “latent” constraints associated with analytically redundant model equations, and “nonlatent” constraints based on causality, and derived from the extended signed directed graph (ESDG). The ESDG is similar to the signed directed graph (SDG) (Iri et al., 1979), but includes certain nonphysical branches that account for complex dynamics (inverse and compensatory response due to negative feedback). Necessary conditions based on the topology of the SDG for the occurrence of complex dynamics are developed. Based on these criteria, assumptions concerning disturbance propagation are introduced that minimize spurious interpretations of system behavior while guaranteeing inclusion of the steady-state response. The method is demonstrated by simulating the effects of equipment malfunctions on a model process.

Ancillary