 |
Center for Nonlinear Studies |
Optimization of thermal fluid and reactive systems have been important tools for engineering process systems. In recent years, Equation oriented (EO) approaches to formulate and solve nonlinear programs (NLPs) for large-scale process engineering systems have led to significant improvements in design, operation and reliability of these systems. Enabling tools are largescale NLP solvers as well as optimization modeling platforms. Recent advances in this area will be discussed at three levels. First, modifications of fast interior point NLP solvers will be discussed and demonstrated on both well-posed and ill-posed optimization problems, including NLPs with complementarity constraints. Second, the formulation of well-posed engineering models will be considered in order to link them to convergence characteristics of the optimization solver. Finally, a number of case studies will be presented for chemical reactive and polymer processes, energy and power systems, and on-line solution of nonlinear dynamic systems. The synergy of these three levels will be demonstrated on real-world industrial examples as well.
Host: Russell Bent