Process Optimization

Session Chairs:

• Armin Fricke, Chemstations Europe 

Session Description:

Process Optimization is undergoing a paradigm shift. In the past, good knowledge about the process, its components, and clear objectives for the optimization were a sufficient starting point for working with a process simulator. Today, process optimization comes from many different directions: Modular plants, autonomous plants, smart maintenance, surrogate modeling, metaheuristics, and cluster computing.

New enablers for process optimization are digitization, which results in the Digital Twin and collaborative workflows, and machine learning, which allows to build gray-box models combining rigorous thermodynamic simulation and plant data.

This session focuses on the methods and tools available for process optimization.

*All session and speaker information is subject to change pending finalization

Confirmed Speakers:

• Norbert Asprion, BASF
• Michael Bortz, Fraunhofer Institute for Industrial Mathematics
• Mathieu Cura, Optimistik
• Kevin McBride, Max Planck Institute for Dynamics of Complex Technical Systems    

Abstracts:

Machine Learning Meets Process Simulator: Greybox and Surrogate Modeling to Support Process Optimization

Norbert Asprion, BASF 

Besides experimentation process modeling, simulation and optimization is one of the key elements of process development. Especially, when comparing process concepts with regard to partly competing sustainability criteria, multicriteria optimization and decision support (cf. for example Bortz et al., 2014, Burger et al., 2014) are indispensable. The applicability of such methods is sometimes restricted either due to missing models for certain process steps to complete a simulation model for the whole process or by costly and hardly solvable simulation models. In both cases machine learning can provide a feasible solution to tackle these restrictions.

For missing scientific process models available process data can be used to set-up data-driven models. In most cases process data are not or only partly available at the interfaces of the process step to be modeled. Therefore, as a first step usually a simplified model for the process step is included in the simulation model to complete the model for the whole process and used as a soft-sensor for the missing information. After this step the necessary information is available to model the process step for example with machine learning methods separately (so-called incremental model identification, cf. e.g. Bardow and Marquardt, 2004) and include it afterwards in the entire process simulation model.

In cases where a process model is too hard to be applied directly for optimization, the use of surrogate modeling (Bhosekar and Ierapetritou, 2018) can be a resort. Here the original model is sampled, and the collected data of these samples can be taken to train for example artificial neural networks reflecting the optimization variables as inputs and the objectives and constraints as outputs. The obtained optimized variables can be used as inputs for the original process model to verify the results of the surrogate model. If the agreement with the predicted results of the surrogate is not satisfying the verification results can be used to retrain the surrogate, run again the optimization and repeat the verification step. This iterative procedure could be repeated until sufficient agreement is obtained.

Both methods will be demonstrated for a process simulation example of a Cumene process.

   N. Asprion, R. Böttcher, M. Stavrou, J. Höller, J. Schwientek, M. Bortz

A. Bardow, W. Marquardt, 2004, Chem. Eng. Sci., 59, 2673 – 2684.

M. Bortz, J. Burger, N. Asprion, S. Blagov, R. Böttcher, U. Nowak, A. Scheithauer, R. Welke, K.-H. Küfer, and H. Hasse, 2014, Comp. Chem. Eng., 60, 354–363. doi:10.1016/j.compchemeng.2013.09.015

J. Burger, N. Asprion, S. Blagov, R. Böttcher, U. Nowak, M. Bortz, R. Welke, K.-H. Küfer, and H. Hasse, 2014, Chem. Ing. Tech., 86 (7), 1065–1072. DOI: 10.1002/cite.201400008

A. Bhosekar, M. Ierapetritou, 2018,Comp. Chem. Eng. 108, 250 – 267.

Multi-Criteria Optimization in Process Design Supported by Machine Learning Methods

Michael Bortz, Fraunhofer Institute for Industrial Mathematics 

M. Bortz, R. Heese, J. Schwientek, R. Böttcher, N. Asprion

Process design is a multicriteria optimization problem: Costs as small as possible, quality indicators as high as possible shall be achieved. This requires calculating the Pareto boundary, which makes many model evaluations necessary and thus is a computationally demanding task. Once the Pareto boundary has been determined, it can be used for an interactive decision support, which enables the engineer to balance the objectives e.g. increase quality at competitive total costs.

In this contribution, we review computationally highly efficient methods to resolve the Pareto boundary within a predefined accuracy. Among these are adaptive scalarization (Bortz et al, 2014) and exploration schemes (Heese et al, 2019), where the latter employ machine-learning methods.  Furthermore, the application of multicriteria approaches in the presence of parametric model uncertainties is described (Bortz et al, 2017).

The benefit of applying multicriteria optimization in process design is reported for different industrial applications. The transferability of the method to data reconciliation and the detection of faulty measurements is demonstrated as well.

M. Bortz, J. Burger, N. Asprion, S. Blagov, R. Böttcher, U. Nowak, A. Scheithauer, R. Welke, K.-H. Küfer, and H. Hasse, 2014, Comp. Chem. Eng., 60, 354–363

M. Bortz, J. Burger, E. von Harbou, M. Klein, J. Schwientek, N. Asprion, R. Böttcher,  K.-H. Küfer, and H. Hasse. Industrial & Engineering Chemistry Research, 56(44):12672-12681, 2017

R. Heese, M. Walczak, T. Seidel, N. Asprion, M. Bortz, 2019, Comp. Chem. Eng. 124, 326-342