(190k) Single and Multi-Objective Optimizations Using Parallelized Process Simulators

Equation-oriented (EO) optimization methods are difficult to apply to rigorous thermodynamic methods, discontinuous variables, or systems where the number of equations is not constant. Despite these challenges, EO methods are still quite commonly used for process optimization, due in part to a perception that EO methods are faster than process simulation-based methods.
We have found that by integrating parallel processing techniques into modern process simulation and optimization, we can achieve rapid results comparable to the speed of more traditional EO modeling methods. We can attain these speeds even while using first-principle models, without risking validity of the model through too many simplifying assumptions. In this presentation, we will consider several process examples to explore the advantages of using simulators for optimization projects.