Simultaneous parameter estimation in reactive-solvent-based processes

Abstract

Accurate estimation of model parameters is difficult for reactive solvent-based processes due to coupled effects of mass transfer, heat transfer, and reaction kinetics. While commercial process simulators provide capabilities for parameter estimation, it is generally difficult to simultaneously regress parameters of different submodels (e.g. parameters of physical properties and mass transfer models). Morgan et al. (2018) and Chinen et al. (2018) have demonstrated the value of the simultaneous parameter estimation approach in developing highly predictive models. The approach used by Chinen et al. (2018) embedded a simulation of a CO2 capture process in an external derivative free optimization (DFO) framework. While this approach works well in many cases, it is computationally inefficient, difficult to set up, and limited by the amount of data that can be used. This paper presents an extension to that work where the process model and a generalized parameter estimation tool are developed in the Institute for Design of Advanced Energy Systems (IDAES) process systems engineering (PSE) framework. The IDAES framework allows an entire process model to be implemented in an equation-oriented framework with access to state-of-the-art optimization methods. Improved computational efficiency and ease of implementation are demonstrated.