Dissipativity based analysis and control of process networks with varying time delay

Abstract

An approach to distributed plant-wide analysis and control of networked systems with varying time delay in discrete time is presented, with a particular focus on chemical process systems. Plant-wide stability and performance conditions are ensured by putting constraints on the dissipativity properties of the plant-wide system, which is a linear function of the dissipativity properties of individual subsystems. To facilitate this, dynamic supply rates in the form of quadratic difference forms are used, which provide less conservative stability conditions than traditional forms of dissipativity and allow for the effect of time delays to be captured. The control design procedure includes a global planning stage where the required dissipativity of each controller is determined, followed by the independent synthesis of each controller which may take the form of either a linear controller or model predictive controller. This framework may be extended to consider additional forms of uncertainty.