Order reduction of large scale DAE models

Authors:	Edgar Thomas, The University of Texas at Austin, United States Hedengren John, The University of Texas at Austin, United States
Topic:	5.4 Large Scale Complex Systems
Session:	Large Scale Complex Systems I- Theory
Keywords:	Order reduction, Model reduction, Nonlinear systems, Function approximation, Distillation columns, Dynamic models

Abstract

This report outlines a technique for the order reduction of differential algebraic equations (DAEs). The order reduction is accomplished in two steps. First, algebraic states are partitioned into successive implicit sets of variables and equations by reconstructing the sparsity pattern into lower triangular block form. Second, proper orthogonal decomposition (POD) is used to reduce the number of differential states. As a test case for the theory, a dynamic binary distillation column model is analyzed with the generalized approach. The index 1 DAE model of 52 differential and 178 algebraic states is reduced to an ODE model of 26 differential states.