Chapter 2 - Don’t search—Solve! Process optimization modeling with IDAES

Abstract

The last 50years have seen tremendous advances in mathematical programming algorithms and software for process optimization. Moreover, powerful optimization modeling environments enable the formulation and solution of large-scale optimization applications. This combination of modern NLP algorithms and optimization platforms has led to fast solution strategies that now routinely solve problems with 104–106 variables, with major impacts in process design, operations, and control. Moreover, mathematical programming algorithms can also integrate with accessible optimization modeling platforms that can be incorporated within a broad spectrum of engineering tasks. This chapter presents and describes the Institute for Design of Advanced Energy Systems Integrated Platform (IDAES), which incorporates all of these concepts within a python-based optimization framework. The platform includes facilities for equation-oriented modeling for static and dynamic processes, exact gradients and Hessians from process models, automated initialization strategies, and seamless interaction with state-of-art large-scale optimization solvers. Finally, a challenging carbon capture case study is presented that demonstrates the significant advantages of this platform.