Energy Systems, Mathematical Modeling(Optimization), Multi-scale Systems, Process Design and Control
As multi-scale modeling and advanced simulation technologies mature, it is desirable to utilize their predictive power in process design and optimization. However, these models present unique challenges for traditional optimization frameworks. These models, such as those for complex fluid flow and molecular dynamics, span a very wide range of length and timescales. The highly specialized methods required to simulate these systems are incompatible with popular process optimization methods.
In this project, we are developing a framework for embedding complex simulations in optimization problems. The framework contains two methodologies that build on each other: Îµ-exact models can be used to get close to an optimum and a trust region algorithm that sequentially updates surrogate models. The trust region algorithm is supported by convergence theory to guarantee finding an optimal solution of the truth model. Although our case studies involve the integration of process models and CFD-like continuum models, the use of surrogate models can bridge the gap between other length and timescales. With this optimization framework, insights from the molecular scale can help the decision process at the device and process scale, in turn impacting the decisions at the enterprise scale. We have applied this method to solve flowsheet optimization problems in carbon capture systems. However, this framework could be applied anywhere surrogate models are used, from detailed property calculations to enterprise planning models.
Heat integration is a classic and ever-challenging problem in process systems engineering. Pinch methods remain highly popular and practical for typical heat integration problems, as well as mass exchange networks, work exchange networks, and more. Mathematical programming has also become and indispensable tool to handle the many tradeoffs in these systems. However, the highly combinatorial nature of these problems usually requires models that sacrifice a certain degree of accuracy or computational tractability. In this work, we strive to combine these two approaches by using thermodynamic insight from pinch-based methods to build tractable optimization models. This strategy has been effectively applied to organic Rankine cycle design in refineries and to sub-ambient system integration in the oxycombustion process. The optimization results offered new insights into important trade-offs such as cycle efficiency and heat recovery, while greatly improving efficiency compared to previous studies. As energy costs become more important for competitiveness, these integration methods will become more important throughout the chemical industries.
I am able to teach all core chemical engineering courses, with particular interest in teaching control, design, thermodynamics, and engineering math/programming. I also look forward to the opportunity to develop graduate courses in optimization and mathematical modeling of both dynamic and steady state processes.
Please visit my personal page: www.johneason.me