PROCESS DYNAMIC-MODELS FOR HETEROGENEOUS CHEMICAL REACTORS - AN APPLICATION OF DYNAMIC SINGULARITY THEORY

By analyzing bifurcations from the marginal gain settings of a nonlinear reactor under PI-control, several characteristics of the closed-loop reactor dynamics are revealed via the center manifold projection and normal form techniques of dynamic singularity theory. Of particular practical interests are the effects of set-point error which can often destabilize a nonlinear reactor. It is shown that judicious choice of the set-point developed here can reduce such undesirable effects. The analysis also offers extremely low dimensional nonlinear models which faithfully reproduce all the closed-loop dynamics and are especially accurate near the marginal gains. These reduced models are hence ideal for the design and tuning of reactor controllers.